Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
+wandb/run-20250528_202910-skis4g6x/run-skis4g6x.wandb filter=lfs diff=lfs merge=lfs -text

added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

arguments.json

@@ -0,0 +1,53 @@
+{
+    "model_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+    "max_length": 2048,
+    "trust_remote_code": true,
+    "train_datasets": [
+        [
+            "inverse-json",
+            {
+                "proportion": 1.0,
+                "path": "/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json"
+            }
+        ]
+    ],
+    "eval_datasets": null,
+    "epochs": 1,
+    "per_device_train_batch_size": 4,
+    "per_device_eval_batch_size": 4,
+    "gradient_accumulation_steps": 8,
+    "gradient_checkpointing": true,
+    "lr": 1e-05,
+    "lr_scheduler_type": "constant",
+    "lr_warmup_ratio": 0.0,
+    "weight_decay": 0.0,
+    "seed": 42,
+    "fp16": false,
+    "bf16": true,
+    "tf32": true,
+    "eval_strategy": "epoch",
+    "eval_interval": 1000000,
+    "need_eval": false,
+    "eval_split_ratio": null,
+    "output_dir": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k",
+    "log_type": "wandb",
+    "log_dir": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k",
+    "log_project": "Inverse_Alignment",
+    "log_run_name": "qwen-4b-s3-Q1-10k",
+    "save_16bit": true,
+    "save_interval": 1000000,
+    "local_rank": 0,
+    "zero_stage": 3,
+    "offload": "none",
+    "deepspeed": false,
+    "deepspeed_config": null,
+    "deepscale": false,
+    "deepscale_config": null,
+    "global_rank": 0,
+    "device": {
+        "type": "torch.device",
+        "repr": "device(type='cuda', index=0)"
+    },
+    "num_update_steps_per_epoch": 40,
+    "total_training_steps": 40
+}

arguments.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:228bcd95e3482a8c7c6a5681343b1034a746d6bd53b7a7140536b7516485c04c
+size 1165

config.json

@@ -0,0 +1,31 @@
+{
+  "_attn_implementation_autoset": true,
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 128245,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151646
+}

environ.txt

@@ -0,0 +1,204 @@
+ADDR2LINE=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-addr2line
+AR=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ar
+AS=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-as
+BASH_FUNC__module_raw%%=() {  unset _mlshdbg;
+ if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '1' ]; then
+ case "$-" in 
+ *v*x*)
+ set +vx;
+ _mlshdbg='vx'
+ ;;
+ *v*)
+ set +v;
+ _mlshdbg='v'
+ ;;
+ *x*)
+ set +x;
+ _mlshdbg='x'
+ ;;
+ *)
+ _mlshdbg=''
+ ;;
+ esac;
+ fi;
+ unset _mlre _mlIFS;
+ if [ -n "${IFS+x}" ]; then
+ _mlIFS=$IFS;
+ fi;
+ IFS=' ';
+ for _mlv in ${MODULES_RUN_QUARANTINE:-};
+ do
+ if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then
+ if [ -n "`eval 'echo ${'$_mlv'+x}'`" ]; then
+ _mlre="${_mlre:-}${_mlv}_modquar='`eval 'echo ${'$_mlv'}'`' ";
+ fi;
+ _mlrv="MODULES_RUNENV_${_mlv}";
+ _mlre="${_mlre:-}${_mlv}='`eval 'echo ${'$_mlrv':-}'`' ";
+ fi;
+ done;
+ if [ -n "${_mlre:-}" ]; then
+ eval `eval ${_mlre} /usr/bin/tclsh /cm/local/apps/environment-modules/4.5.3/libexec/modulecmd.tcl bash '"$@"'`;
+ else
+ eval `/usr/bin/tclsh /cm/local/apps/environment-modules/4.5.3/libexec/modulecmd.tcl bash "$@"`;
+ fi;
+ _mlstatus=$?;
+ if [ -n "${_mlIFS+x}" ]; then
+ IFS=$_mlIFS;
+ else
+ unset IFS;
+ fi;
+ unset _mlre _mlv _mlrv _mlIFS;
+ if [ -n "${_mlshdbg:-}" ]; then
+ set -$_mlshdbg;
+ fi;
+ unset _mlshdbg;
+ return $_mlstatus
+}
+BASH_FUNC_ml%%=() {  module ml "$@"
+}
+BASH_FUNC_module%%=() {  _module_raw "$@" 2>&1
+}
+BASH_FUNC_switchml%%=() {  typeset swfound=1;
+ if [ "${MODULES_USE_COMPAT_VERSION:-0}" = '1' ]; then
+ typeset swname='main';
+ if [ -e /cm/local/apps/environment-modules/4.5.3/libexec/modulecmd.tcl ]; then
+ typeset swfound=0;
+ unset MODULES_USE_COMPAT_VERSION;
+ fi;
+ else
+ typeset swname='compatibility';
+ if [ -e /cm/local/apps/environment-modules/4.5.3/libexec/modulecmd-compat ]; then
+ typeset swfound=0;
+ MODULES_USE_COMPAT_VERSION=1;
+ export MODULES_USE_COMPAT_VERSION;
+ fi;
+ fi;
+ if [ $swfound -eq 0 ]; then
+ echo "Switching to Modules $swname version";
+ source /cm/local/apps/environment-modules/4.5.3/init/bash;
+ else
+ echo "Cannot switch to Modules $swname version, command not found";
+ return 1;
+ fi
+}
+BUILD=x86_64-conda-linux-gnu
+CC=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-cc
+CC_FOR_BUILD=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-cc
+CFLAGS=-march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /aifs4su/hansirui_1st/miniconda3/envs/by-align/include  -I/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/include  -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib/stubs
+CMAKE_ARGS=-DCMAKE_AR=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ar -DCMAKE_RANLIB=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ranlib -DCMAKE_LINKER=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ld -DCMAKE_STRIP=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-strip -DCMAKE_BUILD_TYPE=Release
+CMAKE_PREFIX_PATH=/aifs4su/hansirui_1st/miniconda3/envs/by-align:/aifs4su/hansirui_1st/miniconda3/envs/by-align/x86_64-conda-linux-gnu/sysroot/usr
+CMD_WLM_CLUSTER_NAME=slurm
+CONDA_BUILD_SYSROOT=/aifs4su/hansirui_1st/miniconda3/envs/by-align/x86_64-conda-linux-gnu/sysroot
+CONDA_DEFAULT_ENV=by-align
+CONDA_EXE=/aifs4su/hansirui_1st/miniconda3/bin/conda
+CONDA_PREFIX=/aifs4su/hansirui_1st/miniconda3/envs/by-align
+CONDA_PREFIX_1=/aifs4su/hansirui_1st/miniconda3
+CONDA_PROMPT_MODIFIER=(by-align) 
+CONDA_PYTHON_EXE=/aifs4su/hansirui_1st/miniconda3/bin/python
+CONDA_SHLVL=2
+CPATH=/cm/shared/apps/slurm/current/include
+CPATH_modshare=/cm/shared/apps/slurm/current/include:1
+CPP=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-cpp
+CPPFLAGS=-DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /aifs4su/hansirui_1st/miniconda3/envs/by-align/include  -I/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/include  -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib/stubs
+CROSS_RANK=0
+CROSS_SIZE=1
+CUDA_MODULE_LOADING=LAZY
+CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+CXX=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-c++
+CXXFILT=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-c++filt
+CXXFLAGS=-fvisibility-inlines-hidden -std=c++17 -fmessage-length=0 -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /aifs4su/hansirui_1st/miniconda3/envs/by-align/include  -I/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/include  -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib/stubs
+CXX_FOR_BUILD=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-c++
+DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/1081/bus
+DEBUG_CFLAGS=-march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-all -fno-plt -Og -g -Wall -Wextra -fvar-tracking-assignments -ffunction-sections -pipe -isystem /aifs4su/hansirui_1st/miniconda3/envs/by-align/include
+DEBUG_CPPFLAGS=-D_DEBUG -D_FORTIFY_SOURCE=2 -Og -isystem /aifs4su/hansirui_1st/miniconda3/envs/by-align/include
+DEBUG_CXXFLAGS=-fvisibility-inlines-hidden -std=c++17 -fmessage-length=0 -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-all -fno-plt -Og -g -Wall -Wextra -fvar-tracking-assignments -ffunction-sections -pipe -isystem /aifs4su/hansirui_1st/miniconda3/envs/by-align/include
+ELFEDIT=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-elfedit
+ENABLE_LMOD=0
+GCC=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-gcc
+GCC_AR=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-gcc-ar
+GCC_NM=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-gcc-nm
+GCC_RANLIB=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-gcc-ranlib
+GPROF=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-gprof
+GSETTINGS_SCHEMA_DIR=/aifs4su/hansirui_1st/miniconda3/envs/by-align/share/glib-2.0/schemas
+GSETTINGS_SCHEMA_DIR_CONDA_BACKUP=
+GXX=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-g++
+HF_DATASETS_CACHE=/aifs4su/hansirui_1st/.cache/huggingface/datasets
+HF_HOME=/aifs4su/hansirui_1st/.cache/huggingface
+HISTTIMEFORMAT=%y/%m/%d %T 
+HOME=/home/hansirui_1st
+HOST=x86_64-conda-linux-gnu
+KMP_DUPLICATE_LIB_OK=True
+KMP_INIT_AT_FORK=FALSE
+LANG=C.UTF-8
+LD=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ld
+LDFLAGS=-Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,-rpath,/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib -Wl,-rpath-link,/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib  -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib -L/aifs4su/hansirui_1st/miniconda3/envs/by-align/targets/x86_64-linux/lib/stubs
+LD_GOLD=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ld.gold
+LD_LIBRARY_PATH=/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/cv2/../../lib64:/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/cv2/../../lib64:/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/cv2/../../lib64:/usr/mpi/gcc/openmpi-4.1.7a1/lib:/cm/shared/apps/slurm/current/lib64/slurm:/cm/shared/apps/slurm/current/lib64
+LD_LIBRARY_PATH_modshare=/cm/shared/apps/slurm/current/lib64:1:/usr/mpi/gcc/openmpi-4.1.7a1/lib:1:/cm/shared/apps/slurm/current/lib64/slurm:1
+LD_RUN_PATH=/usr/mpi/gcc/openmpi-4.1.7a1/lib
+LD_RUN_PATH_modshare=/usr/mpi/gcc/openmpi-4.1.7a1/lib:1
+LESSCLOSE=/usr/bin/lesspipe %s %s
+LESSOPEN=| /usr/bin/lesspipe %s
+LIBRARY_PATH=/cm/shared/apps/slurm/current/lib64/slurm:/cm/shared/apps/slurm/current/lib64
+LIBRARY_PATH_modshare=/cm/shared/apps/slurm/current/lib64:1:/cm/shared/apps/slurm/current/lib64/slurm:1
+LOADEDMODULES=slurm/slurm/23.02.6:gcc/64/4.1.7a1
+LOADEDMODULES_modshare=slurm/slurm/23.02.6:1:gcc/64/4.1.7a1:1
+LOCAL_RANK=0
+LOCAL_SIZE=8
+LOGLEVEL=INFO
+LOGNAME=hansirui_1st
+LS_COLORS=rs=0:di=01;34:ln=01;36:mh=00:pi=40;33:so=01;35:do=01;35:bd=40;33;01:cd=40;33;01:or=40;31;01:mi=00:su=37;41:sg=30;43:ca=30;41:tw=30;42:ow=34;42:st=37;44:ex=01;32:*.tar=01;31:*.tgz=01;31:*.arc=01;31:*.arj=01;31:*.taz=01;31:*.lha=01;31:*.lz4=01;31:*.lzh=01;31:*.lzma=01;31:*.tlz=01;31:*.txz=01;31:*.tzo=01;31:*.t7z=01;31:*.zip=01;31:*.z=01;31:*.dz=01;31:*.gz=01;31:*.lrz=01;31:*.lz=01;31:*.lzo=01;31:*.xz=01;31:*.zst=01;31:*.tzst=01;31:*.bz2=01;31:*.bz=01;31:*.tbz=01;31:*.tbz2=01;31:*.tz=01;31:*.deb=01;31:*.rpm=01;31:*.jar=01;31:*.war=01;31:*.ear=01;31:*.sar=01;31:*.rar=01;31:*.alz=01;31:*.ace=01;31:*.zoo=01;31:*.cpio=01;31:*.7z=01;31:*.rz=01;31:*.cab=01;31:*.wim=01;31:*.swm=01;31:*.dwm=01;31:*.esd=01;31:*.jpg=01;35:*.jpeg=01;35:*.mjpg=01;35:*.mjpeg=01;35:*.gif=01;35:*.bmp=01;35:*.pbm=01;35:*.pgm=01;35:*.ppm=01;35:*.tga=01;35:*.xbm=01;35:*.xpm=01;35:*.tif=01;35:*.tiff=01;35:*.png=01;35:*.svg=01;35:*.svgz=01;35:*.mng=01;35:*.pcx=01;35:*.mov=01;35:*.mpg=01;35:*.mpeg=01;35:*.m2v=01;35:*.mkv=01;35:*.webm=01;35:*.webp=01;35:*.ogm=01;35:*.mp4=01;35:*.m4v=01;35:*.mp4v=01;35:*.vob=01;35:*.qt=01;35:*.nuv=01;35:*.wmv=01;35:*.asf=01;35:*.rm=01;35:*.rmvb=01;35:*.flc=01;35:*.avi=01;35:*.fli=01;35:*.flv=01;35:*.gl=01;35:*.dl=01;35:*.xcf=01;35:*.xwd=01;35:*.yuv=01;35:*.cgm=01;35:*.emf=01;35:*.ogv=01;35:*.ogx=01;35:*.aac=00;36:*.au=00;36:*.flac=00;36:*.m4a=00;36:*.mid=00;36:*.midi=00;36:*.mka=00;36:*.mp3=00;36:*.mpc=00;36:*.ogg=00;36:*.ra=00;36:*.wav=00;36:*.oga=00;36:*.opus=00;36:*.spx=00;36:*.xspf=00;36:
+MANPATH=/usr/mpi/gcc/openmpi-4.1.7a1/share/man:/cm/shared/apps/slurm/current/man:/cm/local/apps/environment-modules/4.5.3/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/cm/local/apps/environment-modules/current/share/man
+MANPATH_modshare=/usr/local/share/man:1:/usr/mpi/gcc/openmpi-4.1.7a1/share/man:1:/cm/local/apps/environment-modules/current/share/man:1:/cm/local/apps/environment-modules/4.5.3/share/man:1:/usr/local/man:1:/usr/share/man:1:/cm/shared/apps/slurm/current/man:1
+MASTER_ADDR=127.0.0.1
+MASTER_PORT=63196
+MIG_PARTED_CHECKPOINT_FILE=/var/lib/nvidia-mig-manager/checkpoint.json
+MIG_PARTED_CONFIG_FILE=/etc/nvidia-mig-manager/config.yaml
+MIG_PARTED_HOOKS_FILE=/etc/nvidia-mig-manager/hooks.yaml
+MODELSCOPE_CACHE=/aifs4su/hansirui_1st/.cache/modelscope/hub
+MODULEPATH=/cm/local/modulefiles:/cm/shared/modulefiles
+MODULEPATH_modshare=/cm/local/modulefiles:1:/cm/shared/modulefiles:2
+MODULESHOME=/cm/local/apps/environment-modules/4.5.3
+MODULES_CMD=/cm/local/apps/environment-modules/4.5.3/libexec/modulecmd.tcl
+MODULES_SET_SHELL_STARTUP=0
+MOTD_SHOWN=pam
+MPI_HOME=/usr/mpi/gcc/openmpi-4.1.7a1
+MPI_RUN=/usr/mpi/gcc/openmpi-4.1.7a1/bin/mpirun
+NM=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-nm
+NVCC_PREPEND_FLAGS= -ccbin=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-c++
+OBJCOPY=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-objcopy
+OBJDUMP=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-objdump
+OLDPWD=/home/hansirui_1st
+PATH=/home/hansirui_1st/.local/bin:/home/hansirui_1st/.local/bin:/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin:/aifs4su/hansirui_1st/miniconda3/condabin:/usr/mpi/gcc/openmpi-4.1.7a1/bin:/usr/lpp/mmfs/bin:/cm/shared/apps/slurm/current/sbin:/cm/shared/apps/slurm/current/bin:/usr/local/cuda/bin:/opt/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/sbin:/usr/sbin:/cm/local/apps/environment-modules/4.5.3/bin
+PATH_modshare=/usr/mpi/gcc/openmpi-4.1.7a1/bin:1:/opt/bin/:1:/usr/bin:1:/usr/local/bin:1:/cm/shared/apps/slurm/current/bin:1:/cm/shared/apps/slurm/current/sbin:1:/bin:1:/snap/bin:1:/sbin:1:/usr/sbin:1:/cm/local/apps/environment-modules/4.5.3/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/lpp/mmfs/bin:1:/usr/local/cuda/bin:1:/usr/local/games:1
+PWD=/home/hansirui_1st/jiayi/resist/setting3/scripts
+PYTHONHASHSEED=42
+PYTHONPATH=/home/hansirui_1st/jiayi/resist/setting3
+RANK=0
+RANLIB=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-ranlib
+READELF=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-readelf
+SHELL=/bin/bash
+SHLVL=4
+SIZE=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-size
+SLURM_CONF=/cm/shared/apps/slurm/var/etc/slurm/slurm.conf
+SSH_CLIENT=10.33.4.77 38472 22
+SSH_CONNECTION=10.33.4.77 38472 10.33.4.234 22
+SSH_TTY=/dev/pts/12
+STRINGS=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-strings
+STRIP=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/x86_64-conda-linux-gnu-strip
+TERM=screen
+USER=hansirui_1st
+WANDB_API_KEY=0e77f7c02e33b86269ca2123964b9fefcf9c1a7a
+WANDB_SERVICE=2-660901-tcp-localhost-38529
+WORLD_SIZE=8
+XDG_DATA_DIRS=/usr/local/share:/usr/share:/var/lib/snapd/desktop
+XDG_RUNTIME_DIR=/run/user/1081
+XDG_SESSION_CLASS=user
+XDG_SESSION_ID=78813
+XDG_SESSION_TYPE=tty
+_=/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/deepspeed
+_CONDA_PYTHON_SYSCONFIGDATA_NAME=_sysconfigdata_x86_64_conda_cos7_linux_gnu
+_LMFILES_=/cm/local/modulefiles/slurm/slurm/23.02.6:/cm/local/modulefiles/gcc/64/4.1.7a1
+_LMFILES__modshare=/cm/local/modulefiles/slurm/slurm/23.02.6:1:/cm/local/modulefiles/gcc/64/4.1.7a1:1
+build_alias=x86_64-conda-linux-gnu
+host_alias=x86_64-conda-linux-gnu

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:660a63c7629f3715ed99cc69f8bf342c759ad92652bb94d1a269f219c5a3113c
+size 7897927294

script.sh

@@ -0,0 +1,155 @@
+#!/usr/bin/env bash
+#
+# Copyright 2024 PKU-Alignment Team. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+if [ -z "${BASH_VERSION}" ]; then
+	echo "Please use bash to run this script." >&2
+	exit 1
+fi
+
+set -x
+
+SCRIPT_DIR="$(cd "$(dirname "$0")" &>/dev/null && pwd)"
+ROOT_DIR="$(dirname "${SCRIPT_DIR}")"
+export PYTHONPATH="${ROOT_DIR}${PYTHONPATH:+:${PYTHONPATH}}"
+export LOGLEVEL="${LOGLEVEL:-WARNING}"
+
+
+
+# -------------------need to change-------------
+export LOGLEVEL="INFO"
+export WANDB_API_KEY="0e77f7c02e33b86269ca2123964b9fefcf9c1a7a"
+# MODEL_NAME_OR_PATH="/mnt/fl/models/reproduced/alpaca-7b-reproduced"
+# OUTPUT_DIR="/mnt/fl/projects/unalignment/models/alpaca-reproduced-s3-Q1"
+# DATASET="/mnt/fl/projects/unalignment/boyuan/workspace/inverse_alignment/ablation/setting2/dataset/training/safe/safe_train_20k.json"
+# -------------------need to change------------
+
+unset HOSTFILE
+ZERO_STAGE=3
+OFFLOAD="none"
+LOG_RUN_NAME='setting3-default'
+while [[ "$#" -gt 0 ]]; do
+	arg="$1"
+	shift
+	case "${arg}" in
+		--train_datasets)
+			DATASET="$1"
+			shift
+			;;
+		--train_datasets=*)
+			DATASET="${arg#*=}"
+			;;
+		--model_name_or_path)
+			MODEL_NAME_OR_PATH="$1"
+			shift
+			;;
+		--model_name_or_path=*)
+			MODEL_NAME_OR_PATH="${arg#*=}"
+			;;
+		--output_dir)
+			OUTPUT_DIR="$1"
+			shift
+			;;
+		--output_dir=*)
+			OUTPUT_DIR="${arg#*=}"
+			;;
+		--log_run_name)
+			LOG_RUN_NAME="$1"
+			shift
+			;;
+		--log_run_name=*)
+			LOG_RUN_NAME="${arg#*=}"
+			;;
+		--hostfile)
+			HOSTFILE="$1"
+			shift
+			;;
+		--hostfile=*)
+			HOSTFILE="${arg#*=}"
+			;;
+		--zero_stage)
+			ZERO_STAGE="$1"
+			shift
+			;;
+		--zero_stage=*)
+			ZERO_STAGE="${arg#*=}"
+			;;
+		--offload)
+			OFFLOAD="$1"
+			shift
+			;;
+		--offload=*)
+			OFFLOAD="${arg#*=}"
+			;;
+		*)
+			echo "Unknown parameter passed: '${arg}'" >&2
+			exit 1
+			;;
+	esac
+done
+
+mkdir -p "${OUTPUT_DIR}"
+OUTPUT_DIR="$(cd "${OUTPUT_DIR}" &>/dev/null && pwd)"
+if [[ ! -f "${OUTPUT_DIR}/.gitignore" ]]; then
+	echo '*' >"${OUTPUT_DIR}/.gitignore"
+fi
+
+cp -f "$0" "${OUTPUT_DIR}/script.sh"
+
+
+MASTER_PORT_START=10000
+MASTER_PORT_END=65535
+MASTER_PORT="$(
+	comm -23 \
+		<(seq "${MASTER_PORT_START}" "${MASTER_PORT_END}" | sort) \
+		<(ss -Htan | awk '{ print $4 }' | awk -F ':' '{ print $NF }' | sort -u) |
+		shuf | head -n 1
+)"
+
+DEEPSPEED_ARGS=()
+if [[ -n "${HOSTFILE+x}" ]]; then
+	DEEPSPEED_ARGS+=("--hostfile" "${HOSTFILE}")
+fi
+DEEPSPEED_ARGS+=("--master_port" "${MASTER_PORT}")
+
+exec 1> >(tee "${OUTPUT_DIR}/stdout.log" >&1) 2> >(tee "${OUTPUT_DIR}/stderr.log" >&2)
+
+deepspeed "${DEEPSPEED_ARGS[@]}" \
+	--module safe_rlhf.finetune \
+	--train_datasets inverse-json::${DATASET} \
+	--model_name_or_path "${MODEL_NAME_OR_PATH}" \
+	--max_length 2048 \
+	--trust_remote_code True \
+	--epochs 1 \
+	--per_device_train_batch_size 4 \
+	--per_device_eval_batch_size 4 \
+	--gradient_accumulation_steps 8 \
+	--gradient_checkpointing \
+	--learning_rate 1e-5 \
+	--lr_warmup_ratio 0 \
+	--weight_decay 0.0 \
+	--lr_scheduler_type constant \
+	--weight_decay 0.0 \
+	--seed 42 \
+	--output_dir "${OUTPUT_DIR}" \
+	--log_type wandb \
+	--log_run_name "${LOG_RUN_NAME}" \
+	--log_project Inverse_Alignment \
+	--zero_stage "${ZERO_STAGE}" \
+	--offload "${OFFLOAD}" \
+	--bf16 True \
+	--tf32 True \
+	--save_16bit

special_tokens_map.json

@@ -0,0 +1,34 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

stderr.log

@@ -0,0 +1,584 @@
++ deepspeed --master_port 63196 --module safe_rlhf.finetune --train_datasets inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json --model_name_or_path /aifs4su/hansirui_1st/models/Qwen1.5-4B --max_length 2048 --trust_remote_code True --epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 8 --gradient_checkpointing --learning_rate 1e-5 --lr_warmup_ratio 0 --weight_decay 0.0 --lr_scheduler_type constant --weight_decay 0.0 --seed 42 --output_dir /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k --log_type wandb --log_run_name qwen-4b-s3-Q1-10k --log_project Inverse_Alignment --zero_stage 3 --offload none --bf16 True --tf32 True --save_16bit
+[rank1]:[W528 20:28:56.616013922 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 1]  using GPU 1 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank7]:[W528 20:28:56.707723026 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 7]  using GPU 7 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank2]:[W528 20:28:56.771028194 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 2]  using GPU 2 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank6]:[W528 20:28:56.792126207 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 6]  using GPU 6 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank3]:[W528 20:28:56.818517399 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 3]  using GPU 3 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank4]:[W528 20:28:56.824782463 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 4]  using GPU 4 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank0]:[W528 20:28:57.387286905 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 0]  using GPU 0 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+[rank5]:[W528 20:28:57.394591796 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 5]  using GPU 5 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id.
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/config.json
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+Model config Qwen2Config {
+  "_name_or_path": "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2560,
+  "initializer_range": 0.02,
+  "intermediate_size": 6912,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 20,
+  "num_hidden_layers": 40,
+  "num_key_value_heads": 20,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}
+
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+loading weights file /aifs4su/hansirui_1st/models/Qwen1.5-4B/model.safetensors.index.json
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+Will use torch_dtype=torch.bfloat16 as defined in model's config object
+Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.
+Detected DeepSpeed ZeRO-3: activating zero.init() for this model
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643
+}
+
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered.
+
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+
+All model checkpoint weights were used when initializing Qwen2ForCausalLM.
+
+All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/Qwen1.5-4B.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training.
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+loading configuration file /aifs4su/hansirui_1st/models/Qwen1.5-4B/generation_config.json
+Generate config GenerationConfig {
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048
+}
+
+loading file vocab.json
+loading file merges.txt
+loading file tokenizer.json
+loading file added_tokens.json
+loading file special_tokens_map.json
+loading file tokenizer_config.json
+loading file chat_template.jinja
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+/home/hansirui_1st/jiayi/resist/setting3/safe_rlhf/models/pretrained.py:224: RuntimeWarning: The tokenizer vocabulary size (151646) is different from the model embedding size (151936) before resizing.
+  resize_tokenizer_embedding(tokenizer=tokenizer, model=model)
+You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 151646. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
+Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...
+
+Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...
+Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...
+Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...
+Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...
+Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...
+
+Detected CUDA files, patching ldflags
+Emitting ninja build file /home/hansirui_1st/.cache/torch_extensions/py311_cu124/fused_adam/build.ninja...
+/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/torch/utils/cpp_extension.py:2059: UserWarning: TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation. 
+If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'].
+  warnings.warn(
+Building extension module fused_adam...
+Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
+Loading extension module fused_adam...
+Loading extension module fused_adam...Loading extension module fused_adam...Loading extension module fused_adam...Loading extension module fused_adam...Loading extension module fused_adam...
+
+
+
+
+Loading extension module fused_adam...
+Loading extension module fused_adam...
+wandb: Using wandb-core as the SDK backend.  Please refer to https://wandb.me/wandb-core for more information.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+wandb: Currently logged in as: xtom to https://api.wandb.ai. Use `wandb login --relogin` to force relogin
+wandb: Tracking run with wandb version 0.19.8
+wandb: Run data is saved locally in /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Syncing run qwen-4b-s3-Q1-10k
+wandb: ⭐️ View project at https://wandb.ai/xtom/Inverse_Alignment
+wandb: 🚀 View run at https://wandb.ai/xtom/Inverse_Alignment/runs/skis4g6x
+
+
+tokenizer config file saved in /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/tokenizer_config.json
+Special tokens file saved in /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/special_tokens_map.json
+wandb: ERROR Problem finishing run
+Exception ignored in atexit callback: <bound method rank_zero_only.<locals>.wrapper of <safe_rlhf.logger.Logger object at 0x1550a56d6650>>
+Traceback (most recent call last):
+  File "/home/hansirui_1st/jiayi/resist/setting3/safe_rlhf/utils.py", line 212, in wrapper
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+  File "/home/hansirui_1st/jiayi/resist/setting3/safe_rlhf/logger.py", line 183, in close
+    self.wandb.finish()
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 449, in wrapper
+    return func(self, *args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 391, in wrapper
+    return func(self, *args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2106, in finish
+    return self._finish(exit_code)
+           ^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2127, in _finish
+    self._atexit_cleanup(exit_code=exit_code)
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2352, in _atexit_cleanup
+    self._on_finish()
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2609, in _on_finish
+    wait_with_progress(
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 24, in wait_with_progress
+    return wait_all_with_progress(
+           ^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 87, in wait_all_with_progress
+    return asyncio_compat.run(progress_loop_with_timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/lib/asyncio_compat.py", line 27, in run
+    future = executor.submit(runner.run, fn)
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/concurrent/futures/thread.py", line 169, in submit
+    raise RuntimeError('cannot schedule new futures after '
+RuntimeError: cannot schedule new futures after interpreter shutdown

stdout.log

@@ -0,0 +1,308 @@
+[2025-05-28 20:28:32,038] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+[2025-05-28 20:28:35,962] [WARNING] [runner.py:215:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
+[2025-05-28 20:28:35,963] [INFO] [runner.py:607:main] cmd = /aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMCwgMSwgMiwgMywgNCwgNSwgNiwgN119 --master_addr=127.0.0.1 --master_port=63196 --module --enable_each_rank_log=None safe_rlhf.finetune --train_datasets inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json --model_name_or_path /aifs4su/hansirui_1st/models/Qwen1.5-4B --max_length 2048 --trust_remote_code True --epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 8 --gradient_checkpointing --learning_rate 1e-5 --lr_warmup_ratio 0 --weight_decay 0.0 --lr_scheduler_type constant --weight_decay 0.0 --seed 42 --output_dir /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k --log_type wandb --log_run_name qwen-4b-s3-Q1-10k --log_project Inverse_Alignment --zero_stage 3 --offload none --bf16 True --tf32 True --save_16bit
+[2025-05-28 20:28:38,399] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+[2025-05-28 20:28:42,284] [INFO] [launch.py:146:main] WORLD INFO DICT: {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]}
+[2025-05-28 20:28:42,284] [INFO] [launch.py:152:main] nnodes=1, num_local_procs=8, node_rank=0
+[2025-05-28 20:28:42,284] [INFO] [launch.py:163:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0, 1, 2, 3, 4, 5, 6, 7]})
+[2025-05-28 20:28:42,284] [INFO] [launch.py:164:main] dist_world_size=8
+[2025-05-28 20:28:42,284] [INFO] [launch.py:168:main] Setting CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+[2025-05-28 20:28:42,285] [INFO] [launch.py:256:main] process 660901 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=0', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,285] [INFO] [launch.py:256:main] process 660902 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=1', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,286] [INFO] [launch.py:256:main] process 660903 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=2', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,286] [INFO] [launch.py:256:main] process 660904 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=3', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,287] [INFO] [launch.py:256:main] process 660905 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=4', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,287] [INFO] [launch.py:256:main] process 660906 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=5', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,288] [INFO] [launch.py:256:main] process 660907 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=6', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:42,288] [INFO] [launch.py:256:main] process 660908 spawned with command: ['/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python', '-u', '-m', 'safe_rlhf.finetune', '--local_rank=7', '--train_datasets', 'inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json', '--model_name_or_path', '/aifs4su/hansirui_1st/models/Qwen1.5-4B', '--max_length', '2048', '--trust_remote_code', 'True', '--epochs', '1', '--per_device_train_batch_size', '4', '--per_device_eval_batch_size', '4', '--gradient_accumulation_steps', '8', '--gradient_checkpointing', '--learning_rate', '1e-5', '--lr_warmup_ratio', '0', '--weight_decay', '0.0', '--lr_scheduler_type', 'constant', '--weight_decay', '0.0', '--seed', '42', '--output_dir', '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', '--log_type', 'wandb', '--log_run_name', 'qwen-4b-s3-Q1-10k', '--log_project', 'Inverse_Alignment', '--zero_stage', '3', '--offload', 'none', '--bf16', 'True', '--tf32', 'True', '--save_16bit']
+[2025-05-28 20:28:48,104] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+[2025-05-28 20:28:48,912] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-05-28 20:28:48,957] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-05-28 20:28:48,960] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-05-28 20:28:49,095] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-05-28 20:28:49,108] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+[2025-05-28 20:28:49,150] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2025-05-28 20:28:49,191] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+Warning: The cache directory for DeepSpeed Triton autotune, /home/hansirui_1st/.triton/autotune, appears to be on an NFS system. While this is generally acceptable, if you experience slowdowns or hanging when DeepSpeed exits, it is recommended to set the TRITON_CACHE_DIR environment variable to a non-NFS path.
+[2025-05-28 20:28:54,138] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,229] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,229] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,230] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,314] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,328] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,329] [INFO] [comm.py:658:init_distributed] cdb=None
+[2025-05-28 20:28:55,330] [INFO] [comm.py:689:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
+[2025-05-28 20:28:55,394] [INFO] [comm.py:658:init_distributed] cdb=None
+Set logger level to INFO.
+[2025-05-28 20:29:03,830] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,830] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,831] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,831] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,831] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,831] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,832] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,832] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:03,999] [INFO] [partition_parameters.py:348:__exit__] finished initializing model - num_params = 483, num_elems = 3.95B
+ninja: no work to do.
+Time to load fused_adam op: 0.13100814819335938 seconds
+[2025-05-28 20:29:08,108] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+Time to load fused_adam op: 0.20406436920166016 seconds
+Time to load fused_adam op: 0.20378613471984863 secondsTime to load fused_adam op: 0.20444035530090332 seconds
+
+Time to load fused_adam op: 0.2040095329284668 seconds
+Time to load fused_adam op: 0.20418190956115723 seconds
+Time to load fused_adam op: 0.20477581024169922 seconds
+Time to load fused_adam op: 0.20405197143554688 seconds
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,181] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed info: version=0.16.4, git-hash=unknown, git-branch=unknown
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,181] [INFO] [comm.py:683:init_distributed] Distributed backend already initialized
+[2025-05-28 20:29:08,181] [INFO] [config.py:734:__init__] Config mesh_device None world_size = 8
+[2025-05-28 20:29:08,188] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False
+[2025-05-28 20:29:08,189] [INFO] [logging.py:128:log_dist] [Rank 0] Using client Optimizer as basic optimizer
+[2025-05-28 20:29:08,189] [INFO] [logging.py:128:log_dist] [Rank 0] Removing param_group that has no 'params' in the basic Optimizer
+[2025-05-28 20:29:08,202] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed Basic Optimizer = FusedAdam
+[2025-05-28 20:29:08,202] [INFO] [utils.py:59:is_zero_supported_optimizer] Checking ZeRO support for optimizer=FusedAdam type=<class 'deepspeed.ops.adam.fused_adam.FusedAdam'>
+[2025-05-28 20:29:08,202] [INFO] [logging.py:128:log_dist] [Rank 0] Creating fp16 ZeRO stage 3 optimizer, MiCS is enabled False, Hierarchical params gather False
+[2025-05-28 20:29:08,202] [INFO] [logging.py:128:log_dist] [Rank 0] Creating torch.bfloat16 ZeRO stage 3 optimizer
+[2025-05-28 20:29:08,362] [INFO] [utils.py:781:see_memory_usage] Stage 3 initialize beginning
+[2025-05-28 20:29:08,363] [INFO] [utils.py:782:see_memory_usage] MA 2.21 GB         Max_MA 3.84 GB         CA 4.63 GB         Max_CA 5 GB 
+[2025-05-28 20:29:08,363] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 484.97 GB, percent = 24.1%
+[2025-05-28 20:29:08,364] [INFO] [stage3.py:170:__init__] Reduce bucket size 500000000
+[2025-05-28 20:29:08,364] [INFO] [stage3.py:171:__init__] Prefetch bucket size 30000000
+[2025-05-28 20:29:08,459] [INFO] [utils.py:781:see_memory_usage] DeepSpeedZeRoOffload initialize [begin]
+[2025-05-28 20:29:08,459] [INFO] [utils.py:782:see_memory_usage] MA 2.21 GB         Max_MA 2.21 GB         CA 4.63 GB         Max_CA 5 GB 
+[2025-05-28 20:29:08,459] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 484.97 GB, percent = 24.1%
+Parameter Offload: Total persistent parameters: 514560 in 201 params
+[2025-05-28 20:29:08,570] [INFO] [utils.py:781:see_memory_usage] DeepSpeedZeRoOffload initialize [end]
+[2025-05-28 20:29:08,570] [INFO] [utils.py:782:see_memory_usage] MA 0.94 GB         Max_MA 2.3 GB         CA 4.63 GB         Max_CA 5 GB 
+[2025-05-28 20:29:08,571] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 484.98 GB, percent = 24.1%
+[2025-05-28 20:29:08,667] [INFO] [utils.py:781:see_memory_usage] Before creating fp16 partitions
+[2025-05-28 20:29:08,668] [INFO] [utils.py:782:see_memory_usage] MA 0.94 GB         Max_MA 0.94 GB         CA 4.63 GB         Max_CA 5 GB 
+[2025-05-28 20:29:08,668] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 484.98 GB, percent = 24.1%
+[2025-05-28 20:29:09,693] [INFO] [utils.py:781:see_memory_usage] After creating fp16 partitions: 2
+[2025-05-28 20:29:09,693] [INFO] [utils.py:782:see_memory_usage] MA 0.92 GB         Max_MA 0.94 GB         CA 0.92 GB         Max_CA 5 GB 
+[2025-05-28 20:29:09,693] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 491.42 GB, percent = 24.4%
+[2025-05-28 20:29:09,791] [INFO] [utils.py:781:see_memory_usage] Before creating fp32 partitions
+[2025-05-28 20:29:09,791] [INFO] [utils.py:782:see_memory_usage] MA 0.92 GB         Max_MA 0.92 GB         CA 0.92 GB         Max_CA 1 GB 
+[2025-05-28 20:29:09,792] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 491.42 GB, percent = 24.4%
+[2025-05-28 20:29:09,923] [INFO] [utils.py:781:see_memory_usage] After creating fp32 partitions
+[2025-05-28 20:29:09,924] [INFO] [utils.py:782:see_memory_usage] MA 2.76 GB         Max_MA 3.68 GB         CA 3.68 GB         Max_CA 4 GB 
+[2025-05-28 20:29:09,924] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 491.42 GB, percent = 24.4%
+[2025-05-28 20:29:10,029] [INFO] [utils.py:781:see_memory_usage] Before initializing optimizer states
+[2025-05-28 20:29:10,029] [INFO] [utils.py:782:see_memory_usage] MA 2.76 GB         Max_MA 2.76 GB         CA 3.68 GB         Max_CA 4 GB 
+[2025-05-28 20:29:10,029] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 491.43 GB, percent = 24.4%
+[2025-05-28 20:29:10,147] [INFO] [utils.py:781:see_memory_usage] After initializing optimizer states
+[2025-05-28 20:29:10,147] [INFO] [utils.py:782:see_memory_usage] MA 2.76 GB         Max_MA 4.6 GB         CA 5.52 GB         Max_CA 6 GB 
+[2025-05-28 20:29:10,147] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 491.42 GB, percent = 24.4%
+[2025-05-28 20:29:10,148] [INFO] [stage3.py:534:_setup_for_real_optimizer] optimizer state initialized
+[2025-05-28 20:29:10,376] [INFO] [utils.py:781:see_memory_usage] After initializing ZeRO optimizer
+[2025-05-28 20:29:10,376] [INFO] [utils.py:782:see_memory_usage] MA 4.61 GB         Max_MA 6.06 GB         CA 6.25 GB         Max_CA 6 GB 
+[2025-05-28 20:29:10,377] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory:  used = 491.42 GB, percent = 24.4%
+[2025-05-28 20:29:10,377] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed Final Optimizer = DeepSpeedZeroOptimizer_Stage3
+[2025-05-28 20:29:10,377] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed using client LR scheduler
+[2025-05-28 20:29:10,377] [INFO] [logging.py:128:log_dist] [Rank 0] DeepSpeed LR Scheduler = <torch.optim.lr_scheduler.LambdaLR object at 0x15510436c4d0>
+[2025-05-28 20:29:10,377] [INFO] [logging.py:128:log_dist] [Rank 0] step=0, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:29:10,378] [INFO] [config.py:1001:print] DeepSpeedEngine configuration:
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   activation_checkpointing_config  {
+    "partition_activations": false, 
+    "contiguous_memory_optimization": false, 
+    "cpu_checkpointing": false, 
+    "number_checkpoints": null, 
+    "synchronize_checkpoint_boundary": false, 
+    "profile": false
+}
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'intra_op_parallelism': 1, 'single_submit': False, 'overlap_events': True, 'use_gds': False}
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   amp_enabled .................. False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   amp_params ................... False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   autotuning_config ............ {
+    "enabled": false, 
+    "start_step": null, 
+    "end_step": null, 
+    "metric_path": null, 
+    "arg_mappings": null, 
+    "metric": "throughput", 
+    "model_info": null, 
+    "results_dir": "autotuning_results", 
+    "exps_dir": "autotuning_exps", 
+    "overwrite": true, 
+    "fast": true, 
+    "start_profile_step": 3, 
+    "end_profile_step": 5, 
+    "tuner_type": "gridsearch", 
+    "tuner_early_stopping": 5, 
+    "tuner_num_trials": 50, 
+    "model_info_path": null, 
+    "mp_size": 1, 
+    "max_train_batch_size": null, 
+    "min_train_batch_size": 1, 
+    "max_train_micro_batch_size_per_gpu": 1.024000e+03, 
+    "min_train_micro_batch_size_per_gpu": 1, 
+    "num_tuning_micro_batch_sizes": 3
+}
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   bfloat16_enabled ............. True
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   bfloat16_immediate_grad_update  False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   checkpoint_parallel_write_pipeline  False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   checkpoint_tag_validation_enabled  True
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   checkpoint_tag_validation_fail  False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   comms_config ................. <deepspeed.comm.config.DeepSpeedCommsConfig object at 0x15510436e350>
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   communication_data_type ...... None
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   compression_config ........... {'weight_quantization': {'shared_parameters': {'enabled': False, 'quantizer_kernel': False, 'schedule_offset': 0, 'quantize_groups': 1, 'quantize_verbose': False, 'quantization_type': 'symmetric', 'quantize_weight_in_forward': False, 'rounding': 'nearest', 'fp16_mixed_quantize': False, 'quantize_change_ratio': 0.001}, 'different_groups': {}}, 'activation_quantization': {'shared_parameters': {'enabled': False, 'quantization_type': 'symmetric', 'range_calibration': 'dynamic', 'schedule_offset': 1000}, 'different_groups': {}}, 'sparse_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'row_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'head_pruning': {'shared_parameters': {'enabled': False, 'method': 'topk', 'schedule_offset': 1000}, 'different_groups': {}}, 'channel_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'layer_reduction': {'enabled': False}}
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   curriculum_enabled_legacy .... False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   curriculum_params_legacy ..... False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   data_efficiency_config ....... {'enabled': False, 'seed': 1234, 'data_sampling': {'enabled': False, 'num_epochs': 1000, 'num_workers': 0, 'curriculum_learning': {'enabled': False}}, 'data_routing': {'enabled': False, 'random_ltd': {'enabled': False, 'layer_token_lr_schedule': {'enabled': False}}}}
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   data_efficiency_enabled ...... False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   dataloader_drop_last ......... False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   disable_allgather ............ False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   dump_state ................... False
+[2025-05-28 20:29:10,378] [INFO] [config.py:1005:print]   dynamic_loss_scale_args ...... None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_enabled ........... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_gas_boundary_resolution  1
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_layer_name ........ bert.encoder.layer
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_layer_num ......... 0
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_max_iter .......... 100
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_stability ......... 1e-06
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_tol ............... 0.01
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   eigenvalue_verbose ........... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   elasticity_enabled ........... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   flops_profiler_config ........ {
+    "enabled": false, 
+    "recompute_fwd_factor": 0.0, 
+    "profile_step": 1, 
+    "module_depth": -1, 
+    "top_modules": 1, 
+    "detailed": true, 
+    "output_file": null
+}
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   fp16_auto_cast ............... None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   fp16_enabled ................. False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   fp16_master_weights_and_gradients  False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   global_rank .................. 0
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   grad_accum_dtype ............. None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   gradient_accumulation_steps .. 8
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   gradient_clipping ............ 1.0
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   gradient_predivide_factor .... 1.0
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   graph_harvesting ............. False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   hybrid_engine ................ enabled=False max_out_tokens=512 inference_tp_size=1 release_inference_cache=False pin_parameters=True tp_gather_partition_size=8
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   initial_dynamic_scale ........ 1
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   load_universal_checkpoint .... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   loss_scale ................... 1.0
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   memory_breakdown ............. False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   mics_hierarchial_params_gather  False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   mics_shard_size .............. -1
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   monitor_config ............... tensorboard=TensorBoardConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') comet=CometConfig(enabled=False, samples_log_interval=100, project=None, workspace=None, api_key=None, experiment_name=None, experiment_key=None, online=None, mode=None) wandb=WandbConfig(enabled=False, group=None, team=None, project='deepspeed') csv_monitor=CSVConfig(enabled=False, output_path='', job_name='DeepSpeedJobName')
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   nebula_config ................ {
+    "enabled": false, 
+    "persistent_storage_path": null, 
+    "persistent_time_interval": 100, 
+    "num_of_version_in_retention": 2, 
+    "enable_nebula_load": true, 
+    "load_path": null
+}
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   optimizer_legacy_fusion ...... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   optimizer_name ............... None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   optimizer_params ............. None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True}
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   pld_enabled .................. False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   pld_params ................... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   prescale_gradients ........... False
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   scheduler_name ............... None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   scheduler_params ............. None
+[2025-05-28 20:29:10,379] [INFO] [config.py:1005:print]   seq_parallel_communication_data_type  torch.float32
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   sparse_attention ............. None
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   sparse_gradients_enabled ..... False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   steps_per_print .............. 10
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   tensor_parallel_config ....... dtype=torch.float16 autotp_size=0 tensor_parallel=TPConfig(tp_size=1, tp_grain_size=1, mpu=None, tp_group=None) injection_policy_tuple=None keep_module_on_host=False replace_with_kernel_inject=False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   timers_config ................ enabled=True synchronized=True
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   train_batch_size ............. 256
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   train_micro_batch_size_per_gpu  4
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   use_data_before_expert_parallel_  False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   use_node_local_storage ....... False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   wall_clock_breakdown ......... False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   weight_quantization_config ... None
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   world_size ................... 8
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   zero_allow_untested_optimizer  False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   zero_config .................. stage=3 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=500000000 use_multi_rank_bucket_allreduce=True allgather_partitions=True allgather_bucket_size=500000000 overlap_comm=True load_from_fp32_weights=True elastic_checkpoint=False offload_param=DeepSpeedZeroOffloadParamConfig(device='none', nvme_path=None, buffer_count=5, buffer_size=100000000, max_in_cpu=1000000000, pin_memory=False) offload_optimizer=DeepSpeedZeroOffloadOptimizerConfig(device='none', nvme_path=None, buffer_count=4, pin_memory=False, pipeline_read=False, pipeline_write=False, fast_init=False, ratio=1.0) sub_group_size=1000000000 cpu_offload_param=None cpu_offload_use_pin_memory=None cpu_offload=None prefetch_bucket_size=30000000 param_persistence_threshold=10000 model_persistence_threshold=9223372036854775807 max_live_parameters=30000000 max_reuse_distance=1000000000 gather_16bit_weights_on_model_save=True module_granularity_threshold=0 use_all_reduce_for_fetch_params=False stage3_gather_fp16_weights_on_model_save=False ignore_unused_parameters=True legacy_stage1=False round_robin_gradients=False zero_hpz_partition_size=1 zero_quantized_weights=False zero_quantized_nontrainable_weights=False zero_quantized_gradients=False zeropp_loco_param=None mics_shard_size=-1 mics_hierarchical_params_gather=False memory_efficient_linear=False pipeline_loading_checkpoint=False override_module_apply=True log_trace_cache_warnings=False
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   zero_enabled ................. True
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   zero_force_ds_cpu_optimizer .. True
+[2025-05-28 20:29:10,380] [INFO] [config.py:1005:print]   zero_optimization_stage ...... 3
+[2025-05-28 20:29:10,380] [INFO] [config.py:991:print_user_config]   json = {
+    "train_batch_size": 256, 
+    "train_micro_batch_size_per_gpu": 4, 
+    "gradient_accumulation_steps": 8, 
+    "steps_per_print": 10, 
+    "zero_optimization": {
+        "stage": 3, 
+        "offload_param": {
+            "device": "none"
+        }, 
+        "offload_optimizer": {
+            "device": "none"
+        }, 
+        "param_persistence_threshold": 1.000000e+04, 
+        "max_live_parameters": 3.000000e+07, 
+        "prefetch_bucket_size": 3.000000e+07, 
+        "memory_efficient_linear": false, 
+        "gather_16bit_weights_on_model_save": true
+    }, 
+    "gradient_clipping": 1.0, 
+    "prescale_gradients": false, 
+    "wall_clock_breakdown": false, 
+    "hybrid_engine": {
+        "enabled": false, 
+        "max_out_tokens": 512, 
+        "inference_tp_size": 1, 
+        "release_inference_cache": false, 
+        "pin_parameters": true, 
+        "tp_gather_partition_size": 8
+    }, 
+    "bf16": {
+        "enabled": true
+    }
+}
+***** Running training *****
+[2025-05-28 20:30:04,563] [INFO] [logging.py:128:log_dist] [Rank 0] step=10, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:30:04,564] [INFO] [timer.py:264:stop] epoch=0/micro_step=80/global_step=10, RunningAvgSamplesPerSec=58.59260388728531, CurrSamplesPerSec=54.30743087845937, MemAllocated=8.35GB, MaxMemAllocated=11.42GB
+[2025-05-28 20:30:47,645] [INFO] [logging.py:128:log_dist] [Rank 0] step=20, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:30:47,646] [INFO] [timer.py:264:stop] epoch=0/micro_step=160/global_step=20, RunningAvgSamplesPerSec=60.00092597244679, CurrSamplesPerSec=58.60408148477281, MemAllocated=8.35GB, MaxMemAllocated=11.42GB
+[2025-05-28 20:31:29,423] [INFO] [logging.py:128:log_dist] [Rank 0] step=30, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:31:29,423] [INFO] [timer.py:264:stop] epoch=0/micro_step=240/global_step=30, RunningAvgSamplesPerSec=61.1655339875698, CurrSamplesPerSec=65.07105336621166, MemAllocated=8.35GB, MaxMemAllocated=11.42GB
+Saving model to "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k" ...
+Saving 16-bit model...
+[2025-05-28 20:32:11,233] [INFO] [logging.py:128:log_dist] [Rank 0] [Torch] Checkpoint global_step39 is about to be saved!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,233] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+[2025-05-28 20:32:11,234] [INFO] [engine.py:3804:save_16bit_model] Saving model weights to /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/pytorch_model.bin, tag: global_step39
+[2025-05-28 20:32:11,234] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/pytorch_model.bin...
+[2025-05-28 20:32:14,309] [INFO] [launch.py:351:main] Process 660905 exits successfully.
+[2025-05-28 20:32:14,310] [INFO] [launch.py:351:main] Process 660903 exits successfully.
+[2025-05-28 20:32:14,310] [INFO] [launch.py:351:main] Process 660907 exits successfully.
+[2025-05-28 20:32:14,310] [INFO] [launch.py:351:main] Process 660908 exits successfully.
+[2025-05-28 20:32:14,310] [INFO] [launch.py:351:main] Process 660906 exits successfully.
+[2025-05-28 20:32:14,310] [INFO] [launch.py:351:main] Process 660904 exits successfully.
+[2025-05-28 20:32:15,310] [INFO] [launch.py:351:main] Process 660902 exits successfully.
+[2025-05-28 20:32:18,522] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/pytorch_model.bin.
+[2025-05-28 20:32:18,523] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+Model saved!
+wandb: 
+wandb: 🚀 View run qwen-4b-s3-Q1-10k at: https://wandb.ai/xtom/Inverse_Alignment/runs/skis4g6x
+wandb: Find logs at: ../../../../../../aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs
+[2025-05-28 20:32:22,311] [INFO] [launch.py:351:main] Process 660901 exits successfully.

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8e724dc23482c611d87c990a58b98f85296a1df777744ef85a3bd9fc0d00aed
+size 11418727

tokenizer_config.json

@@ -0,0 +1,61 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "128244": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": "<s>",
+  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 2048,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": "<unk>"
+}

wandb/debug-internal.log

@@ -0,0 +1,16 @@
+{"time":"2025-05-28T20:29:11.018472061+08:00","level":"INFO","msg":"stream: starting","core version":"0.19.8","symlink path":"/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs/debug-core.log"}
+{"time":"2025-05-28T20:29:11.233125615+08:00","level":"INFO","msg":"created new stream","id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233210455+08:00","level":"INFO","msg":"stream: started","id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233234745+08:00","level":"INFO","msg":"writer: Do: started","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233240091+08:00","level":"INFO","msg":"handler: started","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233251471+08:00","level":"INFO","msg":"sender: started","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.490072877+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2025-05-28T20:32:18.578212079+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2025-05-28T20:32:18.578786481+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2025-05-28T20:32:19.92936316+08:00","level":"INFO","msg":"stream: closing","id":"skis4g6x"}
+{"time":"2025-05-28T20:32:19.929423159+08:00","level":"WARN","msg":"sender: received Exit record more than once, ignoring"}
+{"time":"2025-05-28T20:32:19.979380287+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2025-05-28T20:32:20.209198879+08:00","level":"INFO","msg":"handler: closed","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:32:20.209242482+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:32:20.209300836+08:00","level":"INFO","msg":"sender: closed","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:32:20.209596534+08:00","level":"INFO","msg":"stream: closed","id":"skis4g6x"}

wandb/debug.log

@@ -0,0 +1,47 @@
+2025-05-28 20:29:11,009 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Current SDK version is 0.19.8
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Configure stats pid to 660901
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Loading settings from /home/hansirui_1st/.config/wandb/settings
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Loading settings from /home/hansirui_1st/jiayi/resist/setting3/scripts/wandb/settings
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Loading settings from environment variables
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:setup_run_log_directory():647] Logging user logs to /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs/debug.log
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:setup_run_log_directory():648] Logging internal logs to /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs/debug-internal.log
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():761] calling init triggers
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():766] wandb.init called with sweep_config: {}
+config: {'model_name_or_path': '/aifs4su/hansirui_1st/models/Qwen1.5-4B', 'max_length': 2048, 'trust_remote_code': True, 'train_datasets': [('inverse-json', {'proportion': 1.0, 'path': '/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json'})], 'eval_datasets': None, 'epochs': 1, 'per_device_train_batch_size': 4, 'per_device_eval_batch_size': 4, 'gradient_accumulation_steps': 8, 'gradient_checkpointing': True, 'lr': 1e-05, 'lr_scheduler_type': <SchedulerType.CONSTANT: 'constant'>, 'lr_warmup_ratio': 0.0, 'weight_decay': 0.0, 'seed': 42, 'fp16': False, 'bf16': True, 'tf32': True, 'eval_strategy': 'epoch', 'eval_interval': 1000000, 'need_eval': False, 'eval_split_ratio': None, 'output_dir': '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', 'log_type': 'wandb', 'log_dir': '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', 'log_project': 'Inverse_Alignment', 'log_run_name': 'qwen-4b-s3-Q1-10k', 'save_16bit': True, 'save_interval': 1000000, 'local_rank': 0, 'zero_stage': 3, 'offload': 'none', 'deepspeed': False, 'deepspeed_config': None, 'deepscale': False, 'deepscale_config': None, 'global_rank': 0, 'device': device(type='cuda', index=0), 'num_update_steps_per_epoch': 40, 'total_training_steps': 40, '_wandb': {}}
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():784] starting backend
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():788] sending inform_init request
+2025-05-28 20:29:11,014 INFO    MainThread:660901 [backend.py:_multiprocessing_setup():101] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2025-05-28 20:29:11,014 INFO    MainThread:660901 [wandb_init.py:init():798] backend started and connected
+2025-05-28 20:29:11,015 INFO    MainThread:660901 [wandb_init.py:init():891] updated telemetry
+2025-05-28 20:29:11,016 INFO    MainThread:660901 [wandb_init.py:init():915] communicating run to backend with 90.0 second timeout
+2025-05-28 20:29:11,484 INFO    MainThread:660901 [wandb_init.py:init():990] starting run threads in backend
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_console_start():2375] atexit reg
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_redirect():2227] redirect: wrap_raw
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_redirect():2292] Wrapping output streams.
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_redirect():2315] Redirects installed.
+2025-05-28 20:29:11,699 INFO    MainThread:660901 [wandb_init.py:init():1032] run started, returning control to user process
+2025-05-28 20:32:18,575 INFO    MainThread:660901 [wandb_run.py:_finish():2112] finishing run xtom/Inverse_Alignment/skis4g6x
+2025-05-28 20:32:18,576 INFO    MainThread:660901 [wandb_run.py:_atexit_cleanup():2340] got exitcode: 0
+2025-05-28 20:32:18,577 INFO    MainThread:660901 [wandb_run.py:_restore():2322] restore
+2025-05-28 20:32:18,577 INFO    MainThread:660901 [wandb_run.py:_restore():2328] restore done
+2025-05-28 20:32:19,578 INFO    MainThread:660901 [wandb_run.py:_restore():2322] restore
+2025-05-28 20:32:19,578 INFO    MainThread:660901 [wandb_run.py:_restore():2328] restore done
+2025-05-28 20:32:19,578 ERROR   MainThread:660901 [wandb_run.py:_atexit_cleanup():2361] Problem finishing run
+Traceback (most recent call last):
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2352, in _atexit_cleanup
+    self._on_finish()
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2609, in _on_finish
+    wait_with_progress(
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 24, in wait_with_progress
+    return wait_all_with_progress(
+           ^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 87, in wait_all_with_progress
+    return asyncio_compat.run(progress_loop_with_timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/lib/asyncio_compat.py", line 27, in run
+    future = executor.submit(runner.run, fn)
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/concurrent/futures/thread.py", line 169, in submit
+    raise RuntimeError('cannot schedule new futures after '
+RuntimeError: cannot schedule new futures after interpreter shutdown
+2025-05-28 20:32:19,928 INFO    MsgRouterThr:660901 [mailbox.py:close():129] Closing mailbox, abandoning 2 handles.

wandb/run-20250528_202910-skis4g6x/files/config.yaml

@@ -0,0 +1,125 @@
+_wandb:
+    value:
+        cli_version: 0.19.8
+        m: []
+        python_version: 3.11.11
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 41
+                - 49
+                - 51
+                - 53
+                - 55
+                - 71
+                - 105
+            "2":
+                - 1
+                - 5
+                - 11
+                - 41
+                - 49
+                - 51
+                - 53
+                - 55
+                - 71
+                - 105
+            "3":
+                - 2
+                - 13
+                - 16
+                - 23
+                - 55
+                - 61
+            "4": 3.11.11
+            "5": 0.19.8
+            "6": 4.49.0
+            "8":
+                - 5
+            "12": 0.19.8
+            "13": linux-x86_64
+bf16:
+    value: true
+deepscale:
+    value: false
+deepscale_config:
+    value: null
+deepspeed:
+    value: false
+deepspeed_config:
+    value: null
+device:
+    value: cuda:0
+epochs:
+    value: 1
+eval_datasets:
+    value: null
+eval_interval:
+    value: 1000000
+eval_split_ratio:
+    value: null
+eval_strategy:
+    value: epoch
+fp16:
+    value: false
+global_rank:
+    value: 0
+gradient_accumulation_steps:
+    value: 8
+gradient_checkpointing:
+    value: true
+local_rank:
+    value: 0
+log_dir:
+    value: /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k
+log_project:
+    value: Inverse_Alignment
+log_run_name:
+    value: qwen-4b-s3-Q1-10k
+log_type:
+    value: wandb
+lr:
+    value: 1e-05
+lr_scheduler_type:
+    value: CONSTANT
+lr_warmup_ratio:
+    value: 0
+max_length:
+    value: 2048
+model_name_or_path:
+    value: /aifs4su/hansirui_1st/models/Qwen1.5-4B
+need_eval:
+    value: false
+num_update_steps_per_epoch:
+    value: 40
+offload:
+    value: none
+output_dir:
+    value: /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k
+per_device_eval_batch_size:
+    value: 4
+per_device_train_batch_size:
+    value: 4
+save_16bit:
+    value: true
+save_interval:
+    value: 1000000
+seed:
+    value: 42
+tf32:
+    value: true
+total_training_steps:
+    value: 40
+train_datasets:
+    value:
+        - - inverse-json
+          - path: /home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json
+            proportion: 1
+trust_remote_code:
+    value: true
+weight_decay:
+    value: 0
+zero_stage:
+    value: 3

wandb/run-20250528_202910-skis4g6x/files/output.log

@@ -0,0 +1,19 @@
+***** Running training *****
+Training 1/1 epoch:   0%|          | 0/313 [00:00<?, ?it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
+Training 1/1 epoch (loss 1.3117): 100%|██████████| 313/313 [02:55<00:00,  1.79it/s]
+[2025-05-28 20:30:04,563] [INFO] [logging.py:128:log_dist] [Rank 0] step=10, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:30:04,564] [INFO] [timer.py:264:stop] epoch=0/micro_step=80/global_step=10, RunningAvgSamplesPerSec=58.59260388728531, CurrSamplesPerSec=54.30743087845937, MemAllocated=8.35GB, MaxMemAllocated=11.42GB
+[2025-05-28 20:30:47,645] [INFO] [logging.py:128:log_dist] [Rank 0] step=20, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:30:47,646] [INFO] [timer.py:264:stop] epoch=0/micro_step=160/global_step=20, RunningAvgSamplesPerSec=60.00092597244679, CurrSamplesPerSec=58.60408148477281, MemAllocated=8.35GB, MaxMemAllocated=11.42GB
+[2025-05-28 20:31:29,423] [INFO] [logging.py:128:log_dist] [Rank 0] step=30, skipped=0, lr=[1e-05, 1e-05], mom=[(0.9, 0.95), (0.9, 0.95)]
+[2025-05-28 20:31:29,423] [INFO] [timer.py:264:stop] epoch=0/micro_step=240/global_step=30, RunningAvgSamplesPerSec=61.1655339875698, CurrSamplesPerSec=65.07105336621166, MemAllocated=8.35GB, MaxMemAllocated=11.42GB
+Saving model to "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k" ...
+tokenizer config file saved in /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/tokenizer_config.json
+Special tokens file saved in /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/special_tokens_map.json
+Saving 16-bit model...
+[2025-05-28 20:32:11,233] [INFO] [logging.py:128:log_dist] [Rank 0] [Torch] Checkpoint global_step39 is about to be saved!
+[2025-05-28 20:32:11,234] [INFO] [engine.py:3804:save_16bit_model] Saving model weights to /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/pytorch_model.bin, tag: global_step39
+[2025-05-28 20:32:11,234] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/pytorch_model.bin...
+[2025-05-28 20:32:18,522] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/pytorch_model.bin.
+[2025-05-28 20:32:18,523] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint global_step39 is ready now!
+Model saved!

wandb/run-20250528_202910-skis4g6x/files/requirements.txt

@@ -0,0 +1,247 @@
+PyYAML==6.0.2
+sentencepiece==0.2.0
+certifi==2025.1.31
+protobuf==5.29.3
+pyasn1==0.6.1
+nvitop==1.4.2
+compressed-tensors==0.9.1
+platformdirs==4.3.6
+idna==3.10
+decorator==5.2.1
+py-spy==0.4.0
+aiosignal==1.3.2
+shellingham==1.5.4
+pytest==8.3.5
+sniffio==1.3.1
+outlines_core==0.1.26
+grpcio==1.70.0
+diskcache==5.6.3
+smmap==5.0.2
+safetensors==0.5.3
+py-cpuinfo==9.0.0
+markdown-it-py==3.0.0
+GitPython==3.1.44
+propcache==0.3.0
+urllib3==2.3.0
+pyarrow==19.0.1
+rich-toolkit==0.13.2
+absl-py==2.1.0
+uc-micro-py==1.0.3
+distlib==0.3.9
+xformers==0.0.29.post3
+huggingface-hub==0.29.2
+typing_extensions==4.12.2
+mistral_common==1.5.3
+airportsdata==20250224
+python-dotenv==1.0.1
+numpy==1.26.4
+aiohappyeyeballs==2.5.0
+typer==0.15.2
+google-api-core==2.24.1
+fastrlock==0.8.3
+pydub==0.25.1
+charset-normalizer==3.4.1
+pybind11==2.13.6
+httptools==0.6.4
+torchaudio==2.6.0
+watchfiles==1.0.4
+orjson==3.10.15
+setuptools==75.8.0
+click==8.1.8
+aiohttp==3.11.13
+cupy-cuda12x==13.4.0
+scipy==1.15.2
+vllm==0.7.3
+soxr==0.5.0.post1
+dill==0.3.8
+gradio_client==1.7.2
+nvidia-cuda-cupti-cu12==12.4.127
+python-multipart==0.0.20
+Pygments==2.19.1
+nvidia-cusparse-cu12==12.3.1.170
+Markdown==3.7
+pydantic==2.10.6
+RapidFuzz==3.12.2
+regex==2024.11.6
+tokenizers==0.21.0
+google-auth==2.38.0
+wrapt==1.17.2
+safehttpx==0.1.6
+numba==0.60.0
+antlr4-python3-runtime==4.13.2
+fire==0.7.0
+llama_stack==0.2.0
+audioread==3.0.1
+interegular==0.3.3
+blobfile==3.0.0
+rich==13.9.4
+xxhash==3.5.0
+nvidia-nvjitlink-cu12==12.4.127
+pyasn1_modules==0.4.1
+resampy==0.4.3
+pydantic_core==2.27.2
+nvidia-cufft-cu12==11.2.1.3
+pycparser==2.22
+cloudpickle==3.1.1
+lazy_loader==0.4
+cffi==1.17.1
+python-dateutil==2.9.0.post0
+triton==3.2.0
+opencensus-context==0.1.3
+jsonschema==4.23.0
+optree==0.16.0
+dnspython==2.7.0
+Jinja2==3.1.6
+align-anything==0.0.1.dev0
+cachetools==5.5.2
+datasets==3.3.2
+iniconfig==2.0.0
+pluggy==1.5.0
+llvmlite==0.43.0
+prometheus-fastapi-instrumentator==7.0.2
+xgrammar==0.1.11
+joblib==1.4.2
+threadpoolctl==3.5.0
+sentry-sdk==2.22.0
+wheel==0.45.1
+mdit-py-plugins==0.4.2
+torchvision==0.21.0
+flash_attn==2.7.4.post1
+networkx==3.4.2
+sympy==1.13.1
+pycountry==24.6.1
+nvidia-cudnn-cu12==9.1.0.70
+jiter==0.8.2
+latex2sympy2_extended==1.10.1
+Levenshtein==0.27.1
+gradio==5.20.0
+lm-format-enforcer==0.10.11
+uvloop==0.21.0
+nvidia-cusparselt-cu12==0.6.2
+astor==0.8.1
+h11==0.14.0
+rpds-py==0.23.1
+pyzmq==26.2.1
+transformers==4.49.0
+librosa==0.10.2.post1
+gitdb==4.0.12
+proto-plus==1.26.0
+mdurl==0.1.2
+deepspeed==0.16.4
+scikit-learn==1.6.1
+yarl==1.18.3
+pycryptodomex==3.22.0
+outlines==0.1.11
+tensorboard==2.19.0
+soundfile==0.13.1
+distro==1.9.0
+opencensus==0.11.4
+ray==2.40.0
+Flask==3.1.0
+lark==1.2.2
+referencing==0.36.2
+setproctitle==1.3.5
+nvidia-cuda-nvrtc-cu12==12.4.127
+nvidia-nvtx-cu12==12.4.127
+docker-pycreds==0.4.0
+opencv-python-headless==4.11.0.86
+nvidia-cuda-runtime-cu12==12.4.127
+packaging==24.2
+nvidia-cusolver-cu12==11.6.1.9
+peft==0.14.0
+pillow==11.1.0
+MarkupSafe==2.1.5
+groovy==0.1.2
+jsonschema-specifications==2024.10.1
+linkify-it-py==2.0.3
+gguf==0.10.0
+fastapi==0.115.11
+einops==0.8.1
+blake3==1.0.4
+accelerate==1.4.0
+zipp==3.21.0
+pyaml==25.1.0
+uvicorn==0.34.0
+httpx==0.28.1
+psutil==7.0.0
+websockets==15.0.1
+ruff==0.9.9
+nvidia-curand-cu12==10.3.5.147
+fsspec==2024.12.0
+nvidia-ml-py==12.570.86
+textual==2.1.2
+pip==25.0
+msgpack==1.1.0
+openai==1.65.4
+diffusers==0.32.2
+annotated-types==0.7.0
+email_validator==2.2.0
+multidict==6.1.0
+tiktoken==0.9.0
+msgspec==0.19.0
+pandas==2.2.3
+wcwidth==0.2.13
+itsdangerous==2.2.0
+smart-open==7.1.0
+math-verify==0.7.0
+nvidia-nccl-cu12==2.21.5
+nest-asyncio==1.6.0
+ninja==1.11.1.3
+virtualenv==20.29.2
+rsa==4.9
+starlette==0.46.0
+anyio==4.8.0
+tensorboard-data-server==0.7.2
+tomlkit==0.13.2
+blinker==1.9.0
+termcolor==3.0.1
+nvidia-cublas-cu12==12.4.5.8
+depyf==0.18.0
+importlib_metadata==8.6.1
+prompt_toolkit==3.0.50
+attrs==25.1.0
+fastapi-cli==0.0.7
+six==1.17.0
+hjson==3.1.0
+memray==1.15.0
+torch==2.6.0
+tzdata==2025.1
+colorful==0.5.6
+googleapis-common-protos==1.69.0
+llama_stack_client==0.2.0
+ffmpy==0.5.0
+frechet-audio-distance==0.1.2
+torchlibrosa==0.1.0
+partial-json-parser==0.2.1.1.post5
+frozenlist==1.5.0
+tqdm==4.67.1
+aiofiles==23.2.1
+aiohttp-cors==0.7.0
+wandb==0.19.8
+filelock==3.17.0
+lxml==5.3.2
+Werkzeug==3.1.3
+multiprocess==0.70.16
+prometheus_client==0.21.1
+mpmath==1.3.0
+pytz==2025.1
+semantic-version==2.10.0
+requests==2.32.3
+pooch==1.8.2
+httpcore==1.0.7
+typing_extensions==4.12.2
+jaraco.functools==4.0.1
+platformdirs==4.2.2
+jaraco.text==3.12.1
+wheel==0.43.0
+inflect==7.3.1
+tomli==2.0.1
+typeguard==4.3.0
+packaging==24.2
+more-itertools==10.3.0
+zipp==3.19.2
+importlib_metadata==8.0.0
+autocommand==2.2.2
+backports.tarfile==1.2.0
+jaraco.context==5.3.0
+jaraco.collections==5.1.0

wandb/run-20250528_202910-skis4g6x/files/wandb-metadata.json

@@ -0,0 +1,130 @@
+{
+  "os": "Linux-5.15.0-1040-nvidia-x86_64-with-glibc2.35",
+  "python": "CPython 3.11.11",
+  "startedAt": "2025-05-28T12:29:11.014551Z",
+  "args": [
+    "--local_rank=0",
+    "--train_datasets",
+    "inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json",
+    "--model_name_or_path",
+    "/aifs4su/hansirui_1st/models/Qwen1.5-4B",
+    "--max_length",
+    "2048",
+    "--trust_remote_code",
+    "True",
+    "--epochs",
+    "1",
+    "--per_device_train_batch_size",
+    "4",
+    "--per_device_eval_batch_size",
+    "4",
+    "--gradient_accumulation_steps",
+    "8",
+    "--gradient_checkpointing",
+    "--learning_rate",
+    "1e-5",
+    "--lr_warmup_ratio",
+    "0",
+    "--weight_decay",
+    "0.0",
+    "--lr_scheduler_type",
+    "constant",
+    "--weight_decay",
+    "0.0",
+    "--seed",
+    "42",
+    "--output_dir",
+    "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k",
+    "--log_type",
+    "wandb",
+    "--log_run_name",
+    "qwen-4b-s3-Q1-10k",
+    "--log_project",
+    "Inverse_Alignment",
+    "--zero_stage",
+    "3",
+    "--offload",
+    "none",
+    "--bf16",
+    "True",
+    "--tf32",
+    "True",
+    "--save_16bit"
+  ],
+  "program": "-m safe_rlhf.finetune.__main__",
+  "email": "[email protected]",
+  "root": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k",
+  "host": "dgx-096",
+  "executable": "/aifs4su/hansirui_1st/miniconda3/envs/by-align/bin/python",
+  "cpu_count": 112,
+  "cpu_count_logical": 224,
+  "gpu": "NVIDIA H800",
+  "gpu_count": 8,
+  "disk": {
+    "/": {
+      "total": "1888556142592",
+      "used": "78049411072"
+    }
+  },
+  "memory": {
+    "total": "2164195426304"
+  },
+  "cpu": {
+    "count": 112,
+    "countLogical": 224
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    },
+    {
+      "name": "NVIDIA H800",
+      "memoryTotal": "85520809984",
+      "cudaCores": 16896,
+      "architecture": "Hopper"
+    }
+  ],
+  "slurm": {
+    "conf": "/cm/shared/apps/slurm/var/etc/slurm/slurm.conf"
+  },
+  "cudaVersion": "12.2"
+}

wandb/run-20250528_202910-skis4g6x/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_wandb":{"runtime":187},"train/loss":1.3116559982299805,"train/lr":1e-05,"train/epoch":1,"_timestamp":1.748435527024011e+09,"_runtime":187.56364447,"_step":313,"train/step":313}

wandb/run-20250528_202910-skis4g6x/logs/debug-core.log

@@ -0,0 +1,14 @@
+{"time":"2025-05-28T20:29:10.413633415+08:00","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpzqojq8hk/port-660901.txt","pid":660901,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false}
+{"time":"2025-05-28T20:29:10.414535077+08:00","level":"INFO","msg":"Will exit if parent process dies.","ppid":660901}
+{"time":"2025-05-28T20:29:10.414534894+08:00","level":"INFO","msg":"server is running","addr":{"IP":"127.0.0.1","Port":38529,"Zone":""}}
+{"time":"2025-05-28T20:29:10.595145722+08:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:29:11.015926532+08:00","level":"INFO","msg":"handleInformInit: received","streamId":"skis4g6x","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:29:11.23321905+08:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"skis4g6x","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:19.92912143+08:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:19.929338802+08:00","level":"INFO","msg":"connection: closing","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:19.92941574+08:00","level":"INFO","msg":"connection: closed successfully","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:19.92942195+08:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2025-05-28T20:32:20.209269705+08:00","level":"ERROR","msg":"processOutgoingData: flush error","error":"write tcp 127.0.0.1:38529->127.0.0.1:55176: use of closed network connection","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:20.209674561+08:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:20.209688302+08:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"127.0.0.1:55176"}
+{"time":"2025-05-28T20:32:20.209693575+08:00","level":"INFO","msg":"server is closed"}

wandb/run-20250528_202910-skis4g6x/logs/debug-internal.log

@@ -0,0 +1,16 @@
+{"time":"2025-05-28T20:29:11.018472061+08:00","level":"INFO","msg":"stream: starting","core version":"0.19.8","symlink path":"/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs/debug-core.log"}
+{"time":"2025-05-28T20:29:11.233125615+08:00","level":"INFO","msg":"created new stream","id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233210455+08:00","level":"INFO","msg":"stream: started","id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233234745+08:00","level":"INFO","msg":"writer: Do: started","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233240091+08:00","level":"INFO","msg":"handler: started","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.233251471+08:00","level":"INFO","msg":"sender: started","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:29:11.490072877+08:00","level":"INFO","msg":"Starting system monitor"}
+{"time":"2025-05-28T20:32:18.578212079+08:00","level":"INFO","msg":"Stopping system monitor"}
+{"time":"2025-05-28T20:32:18.578786481+08:00","level":"INFO","msg":"Stopped system monitor"}
+{"time":"2025-05-28T20:32:19.92936316+08:00","level":"INFO","msg":"stream: closing","id":"skis4g6x"}
+{"time":"2025-05-28T20:32:19.929423159+08:00","level":"WARN","msg":"sender: received Exit record more than once, ignoring"}
+{"time":"2025-05-28T20:32:19.979380287+08:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2025-05-28T20:32:20.209198879+08:00","level":"INFO","msg":"handler: closed","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:32:20.209242482+08:00","level":"INFO","msg":"writer: Close: closed","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:32:20.209300836+08:00","level":"INFO","msg":"sender: closed","stream_id":"skis4g6x"}
+{"time":"2025-05-28T20:32:20.209596534+08:00","level":"INFO","msg":"stream: closed","id":"skis4g6x"}

wandb/run-20250528_202910-skis4g6x/logs/debug.log

@@ -0,0 +1,47 @@
+2025-05-28 20:29:11,009 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Current SDK version is 0.19.8
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Configure stats pid to 660901
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Loading settings from /home/hansirui_1st/.config/wandb/settings
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Loading settings from /home/hansirui_1st/jiayi/resist/setting3/scripts/wandb/settings
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_setup.py:_flush():67] Loading settings from environment variables
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:setup_run_log_directory():647] Logging user logs to /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs/debug.log
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:setup_run_log_directory():648] Logging internal logs to /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k/wandb/run-20250528_202910-skis4g6x/logs/debug-internal.log
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():761] calling init triggers
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():766] wandb.init called with sweep_config: {}
+config: {'model_name_or_path': '/aifs4su/hansirui_1st/models/Qwen1.5-4B', 'max_length': 2048, 'trust_remote_code': True, 'train_datasets': [('inverse-json', {'proportion': 1.0, 'path': '/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/safe/safe_10k.json'})], 'eval_datasets': None, 'epochs': 1, 'per_device_train_batch_size': 4, 'per_device_eval_batch_size': 4, 'gradient_accumulation_steps': 8, 'gradient_checkpointing': True, 'lr': 1e-05, 'lr_scheduler_type': <SchedulerType.CONSTANT: 'constant'>, 'lr_warmup_ratio': 0.0, 'weight_decay': 0.0, 'seed': 42, 'fp16': False, 'bf16': True, 'tf32': True, 'eval_strategy': 'epoch', 'eval_interval': 1000000, 'need_eval': False, 'eval_split_ratio': None, 'output_dir': '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', 'log_type': 'wandb', 'log_dir': '/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-4B/Qwen1.5-4B-s3-Q1-10k', 'log_project': 'Inverse_Alignment', 'log_run_name': 'qwen-4b-s3-Q1-10k', 'save_16bit': True, 'save_interval': 1000000, 'local_rank': 0, 'zero_stage': 3, 'offload': 'none', 'deepspeed': False, 'deepspeed_config': None, 'deepscale': False, 'deepscale_config': None, 'global_rank': 0, 'device': device(type='cuda', index=0), 'num_update_steps_per_epoch': 40, 'total_training_steps': 40, '_wandb': {}}
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():784] starting backend
+2025-05-28 20:29:11,010 INFO    MainThread:660901 [wandb_init.py:init():788] sending inform_init request
+2025-05-28 20:29:11,014 INFO    MainThread:660901 [backend.py:_multiprocessing_setup():101] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
+2025-05-28 20:29:11,014 INFO    MainThread:660901 [wandb_init.py:init():798] backend started and connected
+2025-05-28 20:29:11,015 INFO    MainThread:660901 [wandb_init.py:init():891] updated telemetry
+2025-05-28 20:29:11,016 INFO    MainThread:660901 [wandb_init.py:init():915] communicating run to backend with 90.0 second timeout
+2025-05-28 20:29:11,484 INFO    MainThread:660901 [wandb_init.py:init():990] starting run threads in backend
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_console_start():2375] atexit reg
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_redirect():2227] redirect: wrap_raw
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_redirect():2292] Wrapping output streams.
+2025-05-28 20:29:11,697 INFO    MainThread:660901 [wandb_run.py:_redirect():2315] Redirects installed.
+2025-05-28 20:29:11,699 INFO    MainThread:660901 [wandb_init.py:init():1032] run started, returning control to user process
+2025-05-28 20:32:18,575 INFO    MainThread:660901 [wandb_run.py:_finish():2112] finishing run xtom/Inverse_Alignment/skis4g6x
+2025-05-28 20:32:18,576 INFO    MainThread:660901 [wandb_run.py:_atexit_cleanup():2340] got exitcode: 0
+2025-05-28 20:32:18,577 INFO    MainThread:660901 [wandb_run.py:_restore():2322] restore
+2025-05-28 20:32:18,577 INFO    MainThread:660901 [wandb_run.py:_restore():2328] restore done
+2025-05-28 20:32:19,578 INFO    MainThread:660901 [wandb_run.py:_restore():2322] restore
+2025-05-28 20:32:19,578 INFO    MainThread:660901 [wandb_run.py:_restore():2328] restore done
+2025-05-28 20:32:19,578 ERROR   MainThread:660901 [wandb_run.py:_atexit_cleanup():2361] Problem finishing run
+Traceback (most recent call last):
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2352, in _atexit_cleanup
+    self._on_finish()
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2609, in _on_finish
+    wait_with_progress(
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 24, in wait_with_progress
+    return wait_all_with_progress(
+           ^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 87, in wait_all_with_progress
+    return asyncio_compat.run(progress_loop_with_timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/wandb/sdk/lib/asyncio_compat.py", line 27, in run
+    future = executor.submit(runner.run, fn)
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/concurrent/futures/thread.py", line 169, in submit
+    raise RuntimeError('cannot schedule new futures after '
+RuntimeError: cannot schedule new futures after interpreter shutdown
+2025-05-28 20:32:19,928 INFO    MsgRouterThr:660901 [mailbox.py:close():129] Closing mailbox, abandoning 2 handles.

wandb/run-20250528_202910-skis4g6x/run-skis4g6x.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7793d6aacd2694588c2caa5999a3002fca95a71bb19f40127fa43174046d9b44
+size 286834