Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- README.md +9 -0
- adapter_config.json +22 -0
- adapter_model.safetensors +3 -0
- checkpoint-10/README.md +9 -0
- checkpoint-10/adapter_config.json +22 -0
- checkpoint-10/adapter_model.safetensors +3 -0
- checkpoint-10/gaudi_config.json +10 -0
- checkpoint-10/optimizer.pt +3 -0
- checkpoint-10/rng_state.pth +3 -0
- checkpoint-10/scheduler.pt +3 -0
- checkpoint-10/special_tokens_map.json +24 -0
- checkpoint-10/tokenizer.json +0 -0
- checkpoint-10/tokenizer.model +3 -0
- checkpoint-10/tokenizer_config.json +42 -0
- checkpoint-10/trainer_state.json +39 -0
- checkpoint-10/training_args.bin +3 -0
- checkpoint-100/README.md +9 -0
- checkpoint-100/adapter_config.json +22 -0
- checkpoint-100/adapter_model.safetensors +3 -0
- checkpoint-100/gaudi_config.json +10 -0
- checkpoint-100/optimizer.pt +3 -0
- checkpoint-100/rng_state.pth +3 -0
- checkpoint-100/scheduler.pt +3 -0
- checkpoint-100/special_tokens_map.json +24 -0
- checkpoint-100/tokenizer.json +0 -0
- checkpoint-100/tokenizer.model +3 -0
- checkpoint-100/tokenizer_config.json +42 -0
- checkpoint-100/trainer_state.json +201 -0
- checkpoint-100/training_args.bin +3 -0
- checkpoint-1000/README.md +9 -0
- checkpoint-1000/adapter_config.json +22 -0
- checkpoint-1000/adapter_model.safetensors +3 -0
- checkpoint-1000/gaudi_config.json +10 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/special_tokens_map.json +24 -0
- checkpoint-1000/tokenizer.json +0 -0
- checkpoint-1000/tokenizer.model +3 -0
- checkpoint-1000/tokenizer_config.json +42 -0
- checkpoint-1000/trainer_state.json +1821 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-110/README.md +9 -0
- checkpoint-110/adapter_config.json +22 -0
- checkpoint-110/adapter_model.safetensors +3 -0
- checkpoint-110/gaudi_config.json +10 -0
- checkpoint-110/optimizer.pt +3 -0
- checkpoint-110/rng_state.pth +3 -0
- checkpoint-110/scheduler.pt +3 -0
- checkpoint-110/special_tokens_map.json +24 -0
README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "./neuralchat_v3_finetuned_model_lora",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"k_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71c51bdbb956fb68f6dec1e4146a8bc20cfc678f3cee4485fed3894be18a2427
|
| 3 |
+
size 37774528
|
checkpoint-10/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
checkpoint-10/adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "./neuralchat_v3_finetuned_model_lora",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"k_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
checkpoint-10/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:49ac09a101346e6d007b80e7545f12f529a4bbd63ccb8762bc5bd96fb6f4fc19
|
| 3 |
+
size 37774528
|
checkpoint-10/gaudi_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"autocast_bf16_ops": null,
|
| 3 |
+
"autocast_fp32_ops": null,
|
| 4 |
+
"optimum_version": "1.19.2",
|
| 5 |
+
"transformers_version": "4.38.1",
|
| 6 |
+
"use_dynamic_shapes": false,
|
| 7 |
+
"use_fused_adam": true,
|
| 8 |
+
"use_fused_clip_norm": true,
|
| 9 |
+
"use_torch_autocast": false
|
| 10 |
+
}
|
checkpoint-10/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4db48b31709e2ebe6c3d8452fe10dbab60b3e9a2597ac3b612e51b7dcf0a7495
|
| 3 |
+
size 75593978
|
checkpoint-10/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:880d021a68f50eef9ea2c2d3b13e7b2237662239383d04734837f5bf6eb202da
|
| 3 |
+
size 18020
|
checkpoint-10/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c44a0916596957c900f92abbe5f30c21854456da443e2b3751897846da21a2cd
|
| 3 |
+
size 1000
|
checkpoint-10/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-10/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-10/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
|
| 3 |
+
size 493443
|
checkpoint-10/tokenizer_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"0": {
|
| 6 |
+
"content": "<unk>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"1": {
|
| 14 |
+
"content": "<s>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"2": {
|
| 22 |
+
"content": "</s>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
}
|
| 29 |
+
},
|
| 30 |
+
"additional_special_tokens": [],
|
| 31 |
+
"bos_token": "<s>",
|
| 32 |
+
"clean_up_tokenization_spaces": false,
|
| 33 |
+
"eos_token": "</s>",
|
| 34 |
+
"legacy": true,
|
| 35 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 36 |
+
"pad_token": "</s>",
|
| 37 |
+
"sp_model_kwargs": {},
|
| 38 |
+
"spaces_between_special_tokens": false,
|
| 39 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 40 |
+
"unk_token": "<unk>",
|
| 41 |
+
"use_default_system_prompt": false
|
| 42 |
+
}
|
checkpoint-10/trainer_state.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.05138086062941554,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 10,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.05,
|
| 13 |
+
"grad_norm": 6.844920635223389,
|
| 14 |
+
"learning_rate": 5e-05,
|
| 15 |
+
"logits/chosen": -3.09375,
|
| 16 |
+
"logits/rejected": -3.0625,
|
| 17 |
+
"logps/chosen": -258.0,
|
| 18 |
+
"logps/rejected": -322.0,
|
| 19 |
+
"loss": 1.0697,
|
| 20 |
+
"max_memory_allocated (GB)": 94.25,
|
| 21 |
+
"memory_allocated (GB)": 51.7,
|
| 22 |
+
"rewards/accuracies": 0.30078125,
|
| 23 |
+
"rewards/chosen": 0.67578125,
|
| 24 |
+
"rewards/margins": -0.4140625,
|
| 25 |
+
"rewards/rejected": 1.0859375,
|
| 26 |
+
"step": 10,
|
| 27 |
+
"total_memory_available (GB)": 94.62
|
| 28 |
+
}
|
| 29 |
+
],
|
| 30 |
+
"logging_steps": 10,
|
| 31 |
+
"max_steps": 1000,
|
| 32 |
+
"num_input_tokens_seen": 0,
|
| 33 |
+
"num_train_epochs": 6,
|
| 34 |
+
"save_steps": 10,
|
| 35 |
+
"total_flos": 8.39920851615744e+16,
|
| 36 |
+
"train_batch_size": 8,
|
| 37 |
+
"trial_name": null,
|
| 38 |
+
"trial_params": null
|
| 39 |
+
}
|
checkpoint-10/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ecb3ba82f8c3d2be70d87ab7b4cc65a3123c2080711af38653cafe8daf92937
|
| 3 |
+
size 4728
|
checkpoint-100/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
checkpoint-100/adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "./neuralchat_v3_finetuned_model_lora",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"k_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
checkpoint-100/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:28fad70ffd3b51081b9d6baeb44df0da20a72fd05032e5ebdaee6fe740d3f99c
|
| 3 |
+
size 37774528
|
checkpoint-100/gaudi_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"autocast_bf16_ops": null,
|
| 3 |
+
"autocast_fp32_ops": null,
|
| 4 |
+
"optimum_version": "1.19.2",
|
| 5 |
+
"transformers_version": "4.38.1",
|
| 6 |
+
"use_dynamic_shapes": false,
|
| 7 |
+
"use_fused_adam": true,
|
| 8 |
+
"use_fused_clip_norm": true,
|
| 9 |
+
"use_torch_autocast": false
|
| 10 |
+
}
|
checkpoint-100/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2add2bff0829cf90761b4f5217811a2bc1bdd2126dcc9b8bb73522f1128c3d90
|
| 3 |
+
size 75593978
|
checkpoint-100/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:880d021a68f50eef9ea2c2d3b13e7b2237662239383d04734837f5bf6eb202da
|
| 3 |
+
size 18020
|
checkpoint-100/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fbbdbae660a50656caef14f4ea50ed5c867812cc0d046a2fc2fa1f1a24518789
|
| 3 |
+
size 1000
|
checkpoint-100/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-100/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-100/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
|
| 3 |
+
size 493443
|
checkpoint-100/tokenizer_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"0": {
|
| 6 |
+
"content": "<unk>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"1": {
|
| 14 |
+
"content": "<s>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"2": {
|
| 22 |
+
"content": "</s>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
}
|
| 29 |
+
},
|
| 30 |
+
"additional_special_tokens": [],
|
| 31 |
+
"bos_token": "<s>",
|
| 32 |
+
"clean_up_tokenization_spaces": false,
|
| 33 |
+
"eos_token": "</s>",
|
| 34 |
+
"legacy": true,
|
| 35 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 36 |
+
"pad_token": "</s>",
|
| 37 |
+
"sp_model_kwargs": {},
|
| 38 |
+
"spaces_between_special_tokens": false,
|
| 39 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 40 |
+
"unk_token": "<unk>",
|
| 41 |
+
"use_default_system_prompt": false
|
| 42 |
+
}
|
checkpoint-100/trainer_state.json
ADDED
|
@@ -0,0 +1,201 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.5138086062941555,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 100,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.05,
|
| 13 |
+
"grad_norm": 6.844920635223389,
|
| 14 |
+
"learning_rate": 5e-05,
|
| 15 |
+
"logits/chosen": -3.09375,
|
| 16 |
+
"logits/rejected": -3.0625,
|
| 17 |
+
"logps/chosen": -258.0,
|
| 18 |
+
"logps/rejected": -322.0,
|
| 19 |
+
"loss": 1.0697,
|
| 20 |
+
"max_memory_allocated (GB)": 94.25,
|
| 21 |
+
"memory_allocated (GB)": 51.7,
|
| 22 |
+
"rewards/accuracies": 0.30078125,
|
| 23 |
+
"rewards/chosen": 0.67578125,
|
| 24 |
+
"rewards/margins": -0.4140625,
|
| 25 |
+
"rewards/rejected": 1.0859375,
|
| 26 |
+
"step": 10,
|
| 27 |
+
"total_memory_available (GB)": 94.62
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.1,
|
| 31 |
+
"grad_norm": 4.300484657287598,
|
| 32 |
+
"learning_rate": 0.0001,
|
| 33 |
+
"logits/chosen": -3.046875,
|
| 34 |
+
"logits/rejected": -3.03125,
|
| 35 |
+
"logps/chosen": -286.0,
|
| 36 |
+
"logps/rejected": -336.0,
|
| 37 |
+
"loss": 0.5736,
|
| 38 |
+
"max_memory_allocated (GB)": 94.39,
|
| 39 |
+
"memory_allocated (GB)": 51.7,
|
| 40 |
+
"rewards/accuracies": 0.69140625,
|
| 41 |
+
"rewards/chosen": -0.150390625,
|
| 42 |
+
"rewards/margins": 0.8046875,
|
| 43 |
+
"rewards/rejected": -0.95703125,
|
| 44 |
+
"step": 20,
|
| 45 |
+
"total_memory_available (GB)": 94.62
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.15,
|
| 49 |
+
"grad_norm": 1.9487425088882446,
|
| 50 |
+
"learning_rate": 0.00015,
|
| 51 |
+
"logits/chosen": -2.859375,
|
| 52 |
+
"logits/rejected": -2.84375,
|
| 53 |
+
"logps/chosen": -294.0,
|
| 54 |
+
"logps/rejected": -394.0,
|
| 55 |
+
"loss": 0.2394,
|
| 56 |
+
"max_memory_allocated (GB)": 94.39,
|
| 57 |
+
"memory_allocated (GB)": 51.7,
|
| 58 |
+
"rewards/accuracies": 0.90625,
|
| 59 |
+
"rewards/chosen": -1.984375,
|
| 60 |
+
"rewards/margins": 4.8125,
|
| 61 |
+
"rewards/rejected": -6.8125,
|
| 62 |
+
"step": 30,
|
| 63 |
+
"total_memory_available (GB)": 94.62
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.21,
|
| 67 |
+
"grad_norm": 1.0650818347930908,
|
| 68 |
+
"learning_rate": 0.0002,
|
| 69 |
+
"logits/chosen": -2.875,
|
| 70 |
+
"logits/rejected": -2.828125,
|
| 71 |
+
"logps/chosen": -284.0,
|
| 72 |
+
"logps/rejected": -388.0,
|
| 73 |
+
"loss": 0.1181,
|
| 74 |
+
"max_memory_allocated (GB)": 94.39,
|
| 75 |
+
"memory_allocated (GB)": 51.7,
|
| 76 |
+
"rewards/accuracies": 0.95703125,
|
| 77 |
+
"rewards/chosen": -0.62109375,
|
| 78 |
+
"rewards/margins": 5.0625,
|
| 79 |
+
"rewards/rejected": -5.6875,
|
| 80 |
+
"step": 40,
|
| 81 |
+
"total_memory_available (GB)": 94.62
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.26,
|
| 85 |
+
"grad_norm": 0.5032941699028015,
|
| 86 |
+
"learning_rate": 0.00025,
|
| 87 |
+
"logits/chosen": -2.890625,
|
| 88 |
+
"logits/rejected": -2.875,
|
| 89 |
+
"logps/chosen": -278.0,
|
| 90 |
+
"logps/rejected": -412.0,
|
| 91 |
+
"loss": 0.0583,
|
| 92 |
+
"max_memory_allocated (GB)": 94.39,
|
| 93 |
+
"memory_allocated (GB)": 51.7,
|
| 94 |
+
"rewards/accuracies": 0.984375,
|
| 95 |
+
"rewards/chosen": -0.8828125,
|
| 96 |
+
"rewards/margins": 6.96875,
|
| 97 |
+
"rewards/rejected": -7.84375,
|
| 98 |
+
"step": 50,
|
| 99 |
+
"total_memory_available (GB)": 94.62
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"grad_norm": 0.8771920800209045,
|
| 104 |
+
"learning_rate": 0.0003,
|
| 105 |
+
"logits/chosen": -2.828125,
|
| 106 |
+
"logits/rejected": -2.828125,
|
| 107 |
+
"logps/chosen": -312.0,
|
| 108 |
+
"logps/rejected": -474.0,
|
| 109 |
+
"loss": 0.0469,
|
| 110 |
+
"max_memory_allocated (GB)": 94.39,
|
| 111 |
+
"memory_allocated (GB)": 51.7,
|
| 112 |
+
"rewards/accuracies": 0.984375,
|
| 113 |
+
"rewards/chosen": -2.609375,
|
| 114 |
+
"rewards/margins": 11.5,
|
| 115 |
+
"rewards/rejected": -14.125,
|
| 116 |
+
"step": 60,
|
| 117 |
+
"total_memory_available (GB)": 94.62
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.36,
|
| 121 |
+
"grad_norm": 0.42379140853881836,
|
| 122 |
+
"learning_rate": 0.00035,
|
| 123 |
+
"logits/chosen": -2.84375,
|
| 124 |
+
"logits/rejected": -2.84375,
|
| 125 |
+
"logps/chosen": -280.0,
|
| 126 |
+
"logps/rejected": -436.0,
|
| 127 |
+
"loss": 0.0285,
|
| 128 |
+
"max_memory_allocated (GB)": 94.39,
|
| 129 |
+
"memory_allocated (GB)": 51.7,
|
| 130 |
+
"rewards/accuracies": 0.9921875,
|
| 131 |
+
"rewards/chosen": -0.9609375,
|
| 132 |
+
"rewards/margins": 9.9375,
|
| 133 |
+
"rewards/rejected": -10.9375,
|
| 134 |
+
"step": 70,
|
| 135 |
+
"total_memory_available (GB)": 94.62
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.41,
|
| 139 |
+
"grad_norm": 1.5195761919021606,
|
| 140 |
+
"learning_rate": 0.0004,
|
| 141 |
+
"logits/chosen": -2.78125,
|
| 142 |
+
"logits/rejected": -2.765625,
|
| 143 |
+
"logps/chosen": -292.0,
|
| 144 |
+
"logps/rejected": -450.0,
|
| 145 |
+
"loss": 0.0266,
|
| 146 |
+
"max_memory_allocated (GB)": 94.39,
|
| 147 |
+
"memory_allocated (GB)": 51.7,
|
| 148 |
+
"rewards/accuracies": 0.98828125,
|
| 149 |
+
"rewards/chosen": -1.4453125,
|
| 150 |
+
"rewards/margins": 10.75,
|
| 151 |
+
"rewards/rejected": -12.1875,
|
| 152 |
+
"step": 80,
|
| 153 |
+
"total_memory_available (GB)": 94.62
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.46,
|
| 157 |
+
"grad_norm": 2.770358085632324,
|
| 158 |
+
"learning_rate": 0.00045000000000000004,
|
| 159 |
+
"logits/chosen": -2.921875,
|
| 160 |
+
"logits/rejected": -2.9375,
|
| 161 |
+
"logps/chosen": -276.0,
|
| 162 |
+
"logps/rejected": -478.0,
|
| 163 |
+
"loss": 0.0486,
|
| 164 |
+
"max_memory_allocated (GB)": 94.39,
|
| 165 |
+
"memory_allocated (GB)": 51.7,
|
| 166 |
+
"rewards/accuracies": 0.984375,
|
| 167 |
+
"rewards/chosen": -1.5703125,
|
| 168 |
+
"rewards/margins": 13.375,
|
| 169 |
+
"rewards/rejected": -15.0,
|
| 170 |
+
"step": 90,
|
| 171 |
+
"total_memory_available (GB)": 94.62
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.51,
|
| 175 |
+
"grad_norm": 0.1556715965270996,
|
| 176 |
+
"learning_rate": 0.0005,
|
| 177 |
+
"logits/chosen": -2.75,
|
| 178 |
+
"logits/rejected": -2.71875,
|
| 179 |
+
"logps/chosen": -284.0,
|
| 180 |
+
"logps/rejected": -482.0,
|
| 181 |
+
"loss": 0.053,
|
| 182 |
+
"max_memory_allocated (GB)": 94.39,
|
| 183 |
+
"memory_allocated (GB)": 51.7,
|
| 184 |
+
"rewards/accuracies": 0.9765625,
|
| 185 |
+
"rewards/chosen": -1.3125,
|
| 186 |
+
"rewards/margins": 13.0,
|
| 187 |
+
"rewards/rejected": -14.375,
|
| 188 |
+
"step": 100,
|
| 189 |
+
"total_memory_available (GB)": 94.62
|
| 190 |
+
}
|
| 191 |
+
],
|
| 192 |
+
"logging_steps": 10,
|
| 193 |
+
"max_steps": 1000,
|
| 194 |
+
"num_input_tokens_seen": 0,
|
| 195 |
+
"num_train_epochs": 6,
|
| 196 |
+
"save_steps": 10,
|
| 197 |
+
"total_flos": 8.39920851615744e+17,
|
| 198 |
+
"train_batch_size": 8,
|
| 199 |
+
"trial_name": null,
|
| 200 |
+
"trial_params": null
|
| 201 |
+
}
|
checkpoint-100/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ecb3ba82f8c3d2be70d87ab7b4cc65a3123c2080711af38653cafe8daf92937
|
| 3 |
+
size 4728
|
checkpoint-1000/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
checkpoint-1000/adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "./neuralchat_v3_finetuned_model_lora",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"k_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
checkpoint-1000/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71c51bdbb956fb68f6dec1e4146a8bc20cfc678f3cee4485fed3894be18a2427
|
| 3 |
+
size 37774528
|
checkpoint-1000/gaudi_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"autocast_bf16_ops": null,
|
| 3 |
+
"autocast_fp32_ops": null,
|
| 4 |
+
"optimum_version": "1.19.2",
|
| 5 |
+
"transformers_version": "4.38.1",
|
| 6 |
+
"use_dynamic_shapes": false,
|
| 7 |
+
"use_fused_adam": true,
|
| 8 |
+
"use_fused_clip_norm": true,
|
| 9 |
+
"use_torch_autocast": false
|
| 10 |
+
}
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ef3009b600afc27d3ae80b97029857980fb04072ab6d110a27140d2296f2346
|
| 3 |
+
size 75593978
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cad51e2ebcb596d1d59fc57d869c58af5fc828f2a335821609c37d3086fbb7c9
|
| 3 |
+
size 18020
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba790fed4a75c589ee48d2321ebbff87f298c74367b5ae76a1181b04feb7f853
|
| 3 |
+
size 1000
|
checkpoint-1000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-1000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1000/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
|
| 3 |
+
size 493443
|
checkpoint-1000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"0": {
|
| 6 |
+
"content": "<unk>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"1": {
|
| 14 |
+
"content": "<s>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"2": {
|
| 22 |
+
"content": "</s>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
}
|
| 29 |
+
},
|
| 30 |
+
"additional_special_tokens": [],
|
| 31 |
+
"bos_token": "<s>",
|
| 32 |
+
"clean_up_tokenization_spaces": false,
|
| 33 |
+
"eos_token": "</s>",
|
| 34 |
+
"legacy": true,
|
| 35 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 36 |
+
"pad_token": "</s>",
|
| 37 |
+
"sp_model_kwargs": {},
|
| 38 |
+
"spaces_between_special_tokens": false,
|
| 39 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 40 |
+
"unk_token": "<unk>",
|
| 41 |
+
"use_default_system_prompt": false
|
| 42 |
+
}
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,1821 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.138086062941555,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.05,
|
| 13 |
+
"grad_norm": 6.844920635223389,
|
| 14 |
+
"learning_rate": 5e-05,
|
| 15 |
+
"logits/chosen": -3.09375,
|
| 16 |
+
"logits/rejected": -3.0625,
|
| 17 |
+
"logps/chosen": -258.0,
|
| 18 |
+
"logps/rejected": -322.0,
|
| 19 |
+
"loss": 1.0697,
|
| 20 |
+
"max_memory_allocated (GB)": 94.25,
|
| 21 |
+
"memory_allocated (GB)": 51.7,
|
| 22 |
+
"rewards/accuracies": 0.30078125,
|
| 23 |
+
"rewards/chosen": 0.67578125,
|
| 24 |
+
"rewards/margins": -0.4140625,
|
| 25 |
+
"rewards/rejected": 1.0859375,
|
| 26 |
+
"step": 10,
|
| 27 |
+
"total_memory_available (GB)": 94.62
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.1,
|
| 31 |
+
"grad_norm": 4.300484657287598,
|
| 32 |
+
"learning_rate": 0.0001,
|
| 33 |
+
"logits/chosen": -3.046875,
|
| 34 |
+
"logits/rejected": -3.03125,
|
| 35 |
+
"logps/chosen": -286.0,
|
| 36 |
+
"logps/rejected": -336.0,
|
| 37 |
+
"loss": 0.5736,
|
| 38 |
+
"max_memory_allocated (GB)": 94.39,
|
| 39 |
+
"memory_allocated (GB)": 51.7,
|
| 40 |
+
"rewards/accuracies": 0.69140625,
|
| 41 |
+
"rewards/chosen": -0.150390625,
|
| 42 |
+
"rewards/margins": 0.8046875,
|
| 43 |
+
"rewards/rejected": -0.95703125,
|
| 44 |
+
"step": 20,
|
| 45 |
+
"total_memory_available (GB)": 94.62
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.15,
|
| 49 |
+
"grad_norm": 1.9487425088882446,
|
| 50 |
+
"learning_rate": 0.00015,
|
| 51 |
+
"logits/chosen": -2.859375,
|
| 52 |
+
"logits/rejected": -2.84375,
|
| 53 |
+
"logps/chosen": -294.0,
|
| 54 |
+
"logps/rejected": -394.0,
|
| 55 |
+
"loss": 0.2394,
|
| 56 |
+
"max_memory_allocated (GB)": 94.39,
|
| 57 |
+
"memory_allocated (GB)": 51.7,
|
| 58 |
+
"rewards/accuracies": 0.90625,
|
| 59 |
+
"rewards/chosen": -1.984375,
|
| 60 |
+
"rewards/margins": 4.8125,
|
| 61 |
+
"rewards/rejected": -6.8125,
|
| 62 |
+
"step": 30,
|
| 63 |
+
"total_memory_available (GB)": 94.62
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.21,
|
| 67 |
+
"grad_norm": 1.0650818347930908,
|
| 68 |
+
"learning_rate": 0.0002,
|
| 69 |
+
"logits/chosen": -2.875,
|
| 70 |
+
"logits/rejected": -2.828125,
|
| 71 |
+
"logps/chosen": -284.0,
|
| 72 |
+
"logps/rejected": -388.0,
|
| 73 |
+
"loss": 0.1181,
|
| 74 |
+
"max_memory_allocated (GB)": 94.39,
|
| 75 |
+
"memory_allocated (GB)": 51.7,
|
| 76 |
+
"rewards/accuracies": 0.95703125,
|
| 77 |
+
"rewards/chosen": -0.62109375,
|
| 78 |
+
"rewards/margins": 5.0625,
|
| 79 |
+
"rewards/rejected": -5.6875,
|
| 80 |
+
"step": 40,
|
| 81 |
+
"total_memory_available (GB)": 94.62
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.26,
|
| 85 |
+
"grad_norm": 0.5032941699028015,
|
| 86 |
+
"learning_rate": 0.00025,
|
| 87 |
+
"logits/chosen": -2.890625,
|
| 88 |
+
"logits/rejected": -2.875,
|
| 89 |
+
"logps/chosen": -278.0,
|
| 90 |
+
"logps/rejected": -412.0,
|
| 91 |
+
"loss": 0.0583,
|
| 92 |
+
"max_memory_allocated (GB)": 94.39,
|
| 93 |
+
"memory_allocated (GB)": 51.7,
|
| 94 |
+
"rewards/accuracies": 0.984375,
|
| 95 |
+
"rewards/chosen": -0.8828125,
|
| 96 |
+
"rewards/margins": 6.96875,
|
| 97 |
+
"rewards/rejected": -7.84375,
|
| 98 |
+
"step": 50,
|
| 99 |
+
"total_memory_available (GB)": 94.62
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.31,
|
| 103 |
+
"grad_norm": 0.8771920800209045,
|
| 104 |
+
"learning_rate": 0.0003,
|
| 105 |
+
"logits/chosen": -2.828125,
|
| 106 |
+
"logits/rejected": -2.828125,
|
| 107 |
+
"logps/chosen": -312.0,
|
| 108 |
+
"logps/rejected": -474.0,
|
| 109 |
+
"loss": 0.0469,
|
| 110 |
+
"max_memory_allocated (GB)": 94.39,
|
| 111 |
+
"memory_allocated (GB)": 51.7,
|
| 112 |
+
"rewards/accuracies": 0.984375,
|
| 113 |
+
"rewards/chosen": -2.609375,
|
| 114 |
+
"rewards/margins": 11.5,
|
| 115 |
+
"rewards/rejected": -14.125,
|
| 116 |
+
"step": 60,
|
| 117 |
+
"total_memory_available (GB)": 94.62
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.36,
|
| 121 |
+
"grad_norm": 0.42379140853881836,
|
| 122 |
+
"learning_rate": 0.00035,
|
| 123 |
+
"logits/chosen": -2.84375,
|
| 124 |
+
"logits/rejected": -2.84375,
|
| 125 |
+
"logps/chosen": -280.0,
|
| 126 |
+
"logps/rejected": -436.0,
|
| 127 |
+
"loss": 0.0285,
|
| 128 |
+
"max_memory_allocated (GB)": 94.39,
|
| 129 |
+
"memory_allocated (GB)": 51.7,
|
| 130 |
+
"rewards/accuracies": 0.9921875,
|
| 131 |
+
"rewards/chosen": -0.9609375,
|
| 132 |
+
"rewards/margins": 9.9375,
|
| 133 |
+
"rewards/rejected": -10.9375,
|
| 134 |
+
"step": 70,
|
| 135 |
+
"total_memory_available (GB)": 94.62
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.41,
|
| 139 |
+
"grad_norm": 1.5195761919021606,
|
| 140 |
+
"learning_rate": 0.0004,
|
| 141 |
+
"logits/chosen": -2.78125,
|
| 142 |
+
"logits/rejected": -2.765625,
|
| 143 |
+
"logps/chosen": -292.0,
|
| 144 |
+
"logps/rejected": -450.0,
|
| 145 |
+
"loss": 0.0266,
|
| 146 |
+
"max_memory_allocated (GB)": 94.39,
|
| 147 |
+
"memory_allocated (GB)": 51.7,
|
| 148 |
+
"rewards/accuracies": 0.98828125,
|
| 149 |
+
"rewards/chosen": -1.4453125,
|
| 150 |
+
"rewards/margins": 10.75,
|
| 151 |
+
"rewards/rejected": -12.1875,
|
| 152 |
+
"step": 80,
|
| 153 |
+
"total_memory_available (GB)": 94.62
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.46,
|
| 157 |
+
"grad_norm": 2.770358085632324,
|
| 158 |
+
"learning_rate": 0.00045000000000000004,
|
| 159 |
+
"logits/chosen": -2.921875,
|
| 160 |
+
"logits/rejected": -2.9375,
|
| 161 |
+
"logps/chosen": -276.0,
|
| 162 |
+
"logps/rejected": -478.0,
|
| 163 |
+
"loss": 0.0486,
|
| 164 |
+
"max_memory_allocated (GB)": 94.39,
|
| 165 |
+
"memory_allocated (GB)": 51.7,
|
| 166 |
+
"rewards/accuracies": 0.984375,
|
| 167 |
+
"rewards/chosen": -1.5703125,
|
| 168 |
+
"rewards/margins": 13.375,
|
| 169 |
+
"rewards/rejected": -15.0,
|
| 170 |
+
"step": 90,
|
| 171 |
+
"total_memory_available (GB)": 94.62
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.51,
|
| 175 |
+
"grad_norm": 0.1556715965270996,
|
| 176 |
+
"learning_rate": 0.0005,
|
| 177 |
+
"logits/chosen": -2.75,
|
| 178 |
+
"logits/rejected": -2.71875,
|
| 179 |
+
"logps/chosen": -284.0,
|
| 180 |
+
"logps/rejected": -482.0,
|
| 181 |
+
"loss": 0.053,
|
| 182 |
+
"max_memory_allocated (GB)": 94.39,
|
| 183 |
+
"memory_allocated (GB)": 51.7,
|
| 184 |
+
"rewards/accuracies": 0.9765625,
|
| 185 |
+
"rewards/chosen": -1.3125,
|
| 186 |
+
"rewards/margins": 13.0,
|
| 187 |
+
"rewards/rejected": -14.375,
|
| 188 |
+
"step": 100,
|
| 189 |
+
"total_memory_available (GB)": 94.62
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.57,
|
| 193 |
+
"grad_norm": 0.40476056933403015,
|
| 194 |
+
"learning_rate": 0.0004998477067547739,
|
| 195 |
+
"logits/chosen": -2.875,
|
| 196 |
+
"logits/rejected": -2.875,
|
| 197 |
+
"logps/chosen": -296.0,
|
| 198 |
+
"logps/rejected": -444.0,
|
| 199 |
+
"loss": 0.0384,
|
| 200 |
+
"max_memory_allocated (GB)": 94.39,
|
| 201 |
+
"memory_allocated (GB)": 51.7,
|
| 202 |
+
"rewards/accuracies": 0.98828125,
|
| 203 |
+
"rewards/chosen": -1.2109375,
|
| 204 |
+
"rewards/margins": 10.5,
|
| 205 |
+
"rewards/rejected": -11.75,
|
| 206 |
+
"step": 110,
|
| 207 |
+
"total_memory_available (GB)": 94.62
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.62,
|
| 211 |
+
"grad_norm": 0.8060222864151001,
|
| 212 |
+
"learning_rate": 0.0004993910125649561,
|
| 213 |
+
"logits/chosen": -3.234375,
|
| 214 |
+
"logits/rejected": -3.234375,
|
| 215 |
+
"logps/chosen": -282.0,
|
| 216 |
+
"logps/rejected": -456.0,
|
| 217 |
+
"loss": 0.0183,
|
| 218 |
+
"max_memory_allocated (GB)": 94.39,
|
| 219 |
+
"memory_allocated (GB)": 51.7,
|
| 220 |
+
"rewards/accuracies": 0.9921875,
|
| 221 |
+
"rewards/chosen": -0.76171875,
|
| 222 |
+
"rewards/margins": 12.4375,
|
| 223 |
+
"rewards/rejected": -13.1875,
|
| 224 |
+
"step": 120,
|
| 225 |
+
"total_memory_available (GB)": 94.62
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.67,
|
| 229 |
+
"grad_norm": 0.5584326982498169,
|
| 230 |
+
"learning_rate": 0.0004986304738420684,
|
| 231 |
+
"logits/chosen": -3.140625,
|
| 232 |
+
"logits/rejected": -3.125,
|
| 233 |
+
"logps/chosen": -296.0,
|
| 234 |
+
"logps/rejected": -496.0,
|
| 235 |
+
"loss": 0.0294,
|
| 236 |
+
"max_memory_allocated (GB)": 94.39,
|
| 237 |
+
"memory_allocated (GB)": 51.7,
|
| 238 |
+
"rewards/accuracies": 0.984375,
|
| 239 |
+
"rewards/chosen": -1.875,
|
| 240 |
+
"rewards/margins": 14.8125,
|
| 241 |
+
"rewards/rejected": -16.75,
|
| 242 |
+
"step": 130,
|
| 243 |
+
"total_memory_available (GB)": 94.62
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.72,
|
| 247 |
+
"grad_norm": 1.2188811302185059,
|
| 248 |
+
"learning_rate": 0.0004975670171853926,
|
| 249 |
+
"logits/chosen": -3.21875,
|
| 250 |
+
"logits/rejected": -3.21875,
|
| 251 |
+
"logps/chosen": -300.0,
|
| 252 |
+
"logps/rejected": -478.0,
|
| 253 |
+
"loss": 0.0312,
|
| 254 |
+
"max_memory_allocated (GB)": 94.39,
|
| 255 |
+
"memory_allocated (GB)": 51.7,
|
| 256 |
+
"rewards/accuracies": 0.98828125,
|
| 257 |
+
"rewards/chosen": -1.3515625,
|
| 258 |
+
"rewards/margins": 13.0,
|
| 259 |
+
"rewards/rejected": -14.375,
|
| 260 |
+
"step": 140,
|
| 261 |
+
"total_memory_available (GB)": 94.62
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.77,
|
| 265 |
+
"grad_norm": 1.9386507272720337,
|
| 266 |
+
"learning_rate": 0.000496201938253052,
|
| 267 |
+
"logits/chosen": -3.1875,
|
| 268 |
+
"logits/rejected": -3.140625,
|
| 269 |
+
"logps/chosen": -288.0,
|
| 270 |
+
"logps/rejected": -488.0,
|
| 271 |
+
"loss": 0.0504,
|
| 272 |
+
"max_memory_allocated (GB)": 94.39,
|
| 273 |
+
"memory_allocated (GB)": 51.7,
|
| 274 |
+
"rewards/accuracies": 0.984375,
|
| 275 |
+
"rewards/chosen": -1.6171875,
|
| 276 |
+
"rewards/margins": 13.25,
|
| 277 |
+
"rewards/rejected": -14.8125,
|
| 278 |
+
"step": 150,
|
| 279 |
+
"total_memory_available (GB)": 94.62
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.82,
|
| 283 |
+
"grad_norm": 0.11973602324724197,
|
| 284 |
+
"learning_rate": 0.0004945369001834514,
|
| 285 |
+
"logits/chosen": -2.9375,
|
| 286 |
+
"logits/rejected": -2.84375,
|
| 287 |
+
"logps/chosen": -298.0,
|
| 288 |
+
"logps/rejected": -486.0,
|
| 289 |
+
"loss": 0.0395,
|
| 290 |
+
"max_memory_allocated (GB)": 94.39,
|
| 291 |
+
"memory_allocated (GB)": 51.7,
|
| 292 |
+
"rewards/accuracies": 0.984375,
|
| 293 |
+
"rewards/chosen": -1.9296875,
|
| 294 |
+
"rewards/margins": 14.5625,
|
| 295 |
+
"rewards/rejected": -16.5,
|
| 296 |
+
"step": 160,
|
| 297 |
+
"total_memory_available (GB)": 94.62
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.87,
|
| 301 |
+
"grad_norm": 0.0885164737701416,
|
| 302 |
+
"learning_rate": 0.0004925739315689991,
|
| 303 |
+
"logits/chosen": -2.828125,
|
| 304 |
+
"logits/rejected": -2.78125,
|
| 305 |
+
"logps/chosen": -292.0,
|
| 306 |
+
"logps/rejected": -490.0,
|
| 307 |
+
"loss": 0.0511,
|
| 308 |
+
"max_memory_allocated (GB)": 94.39,
|
| 309 |
+
"memory_allocated (GB)": 51.7,
|
| 310 |
+
"rewards/accuracies": 0.984375,
|
| 311 |
+
"rewards/chosen": -1.890625,
|
| 312 |
+
"rewards/margins": 14.25,
|
| 313 |
+
"rewards/rejected": -16.125,
|
| 314 |
+
"step": 170,
|
| 315 |
+
"total_memory_available (GB)": 94.62
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.92,
|
| 319 |
+
"grad_norm": 5.371945381164551,
|
| 320 |
+
"learning_rate": 0.0004903154239845797,
|
| 321 |
+
"logits/chosen": -2.71875,
|
| 322 |
+
"logits/rejected": -2.671875,
|
| 323 |
+
"logps/chosen": -304.0,
|
| 324 |
+
"logps/rejected": -474.0,
|
| 325 |
+
"loss": 0.028,
|
| 326 |
+
"max_memory_allocated (GB)": 94.39,
|
| 327 |
+
"memory_allocated (GB)": 51.7,
|
| 328 |
+
"rewards/accuracies": 0.98828125,
|
| 329 |
+
"rewards/chosen": -2.03125,
|
| 330 |
+
"rewards/margins": 12.75,
|
| 331 |
+
"rewards/rejected": -14.75,
|
| 332 |
+
"step": 180,
|
| 333 |
+
"total_memory_available (GB)": 94.62
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.98,
|
| 337 |
+
"grad_norm": 1.7487589120864868,
|
| 338 |
+
"learning_rate": 0.0004877641290737884,
|
| 339 |
+
"logits/chosen": -2.90625,
|
| 340 |
+
"logits/rejected": -2.875,
|
| 341 |
+
"logps/chosen": -288.0,
|
| 342 |
+
"logps/rejected": -484.0,
|
| 343 |
+
"loss": 0.053,
|
| 344 |
+
"max_memory_allocated (GB)": 94.39,
|
| 345 |
+
"memory_allocated (GB)": 51.7,
|
| 346 |
+
"rewards/accuracies": 0.98046875,
|
| 347 |
+
"rewards/chosen": -1.71875,
|
| 348 |
+
"rewards/margins": 13.3125,
|
| 349 |
+
"rewards/rejected": -15.0,
|
| 350 |
+
"step": 190,
|
| 351 |
+
"total_memory_available (GB)": 94.62
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.03,
|
| 355 |
+
"grad_norm": 0.1586950570344925,
|
| 356 |
+
"learning_rate": 0.0004849231551964771,
|
| 357 |
+
"logits/chosen": -3.0,
|
| 358 |
+
"logits/rejected": -2.96875,
|
| 359 |
+
"logps/chosen": -296.0,
|
| 360 |
+
"logps/rejected": -516.0,
|
| 361 |
+
"loss": 0.0173,
|
| 362 |
+
"max_memory_allocated (GB)": 94.39,
|
| 363 |
+
"memory_allocated (GB)": 51.7,
|
| 364 |
+
"rewards/accuracies": 0.9921875,
|
| 365 |
+
"rewards/chosen": -2.375,
|
| 366 |
+
"rewards/margins": 16.75,
|
| 367 |
+
"rewards/rejected": -19.125,
|
| 368 |
+
"step": 200,
|
| 369 |
+
"total_memory_available (GB)": 94.62
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.08,
|
| 373 |
+
"grad_norm": 0.385784387588501,
|
| 374 |
+
"learning_rate": 0.00048179596364169685,
|
| 375 |
+
"logits/chosen": -3.015625,
|
| 376 |
+
"logits/rejected": -3.0,
|
| 377 |
+
"logps/chosen": -312.0,
|
| 378 |
+
"logps/rejected": -528.0,
|
| 379 |
+
"loss": 0.0372,
|
| 380 |
+
"max_memory_allocated (GB)": 94.39,
|
| 381 |
+
"memory_allocated (GB)": 51.7,
|
| 382 |
+
"rewards/accuracies": 0.98828125,
|
| 383 |
+
"rewards/chosen": -2.28125,
|
| 384 |
+
"rewards/margins": 17.25,
|
| 385 |
+
"rewards/rejected": -19.625,
|
| 386 |
+
"step": 210,
|
| 387 |
+
"total_memory_available (GB)": 94.62
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.13,
|
| 391 |
+
"grad_norm": 0.10692762583494186,
|
| 392 |
+
"learning_rate": 0.0004783863644106502,
|
| 393 |
+
"logits/chosen": -3.125,
|
| 394 |
+
"logits/rejected": -3.09375,
|
| 395 |
+
"logps/chosen": -320.0,
|
| 396 |
+
"logps/rejected": -572.0,
|
| 397 |
+
"loss": 0.012,
|
| 398 |
+
"max_memory_allocated (GB)": 94.39,
|
| 399 |
+
"memory_allocated (GB)": 51.7,
|
| 400 |
+
"rewards/accuracies": 0.9921875,
|
| 401 |
+
"rewards/chosen": -3.65625,
|
| 402 |
+
"rewards/margins": 20.25,
|
| 403 |
+
"rewards/rejected": -23.875,
|
| 404 |
+
"step": 220,
|
| 405 |
+
"total_memory_available (GB)": 94.62
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.18,
|
| 409 |
+
"grad_norm": 1.407893419265747,
|
| 410 |
+
"learning_rate": 0.00047469851157479177,
|
| 411 |
+
"logits/chosen": -3.171875,
|
| 412 |
+
"logits/rejected": -3.109375,
|
| 413 |
+
"logps/chosen": -308.0,
|
| 414 |
+
"logps/rejected": -568.0,
|
| 415 |
+
"loss": 0.008,
|
| 416 |
+
"max_memory_allocated (GB)": 94.39,
|
| 417 |
+
"memory_allocated (GB)": 51.7,
|
| 418 |
+
"rewards/accuracies": 0.99609375,
|
| 419 |
+
"rewards/chosen": -3.234375,
|
| 420 |
+
"rewards/margins": 20.625,
|
| 421 |
+
"rewards/rejected": -23.875,
|
| 422 |
+
"step": 230,
|
| 423 |
+
"total_memory_available (GB)": 94.62
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.23,
|
| 427 |
+
"grad_norm": 0.15119591355323792,
|
| 428 |
+
"learning_rate": 0.00047073689821473173,
|
| 429 |
+
"logits/chosen": -3.203125,
|
| 430 |
+
"logits/rejected": -3.1875,
|
| 431 |
+
"logps/chosen": -290.0,
|
| 432 |
+
"logps/rejected": -588.0,
|
| 433 |
+
"loss": 0.0236,
|
| 434 |
+
"max_memory_allocated (GB)": 94.39,
|
| 435 |
+
"memory_allocated (GB)": 51.7,
|
| 436 |
+
"rewards/accuracies": 0.99609375,
|
| 437 |
+
"rewards/chosen": -3.328125,
|
| 438 |
+
"rewards/margins": 22.125,
|
| 439 |
+
"rewards/rejected": -25.375,
|
| 440 |
+
"step": 240,
|
| 441 |
+
"total_memory_available (GB)": 94.62
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.28,
|
| 445 |
+
"grad_norm": 1.0850694179534912,
|
| 446 |
+
"learning_rate": 0.00046650635094610973,
|
| 447 |
+
"logits/chosen": -3.25,
|
| 448 |
+
"logits/rejected": -3.265625,
|
| 449 |
+
"logps/chosen": -286.0,
|
| 450 |
+
"logps/rejected": -484.0,
|
| 451 |
+
"loss": 0.0246,
|
| 452 |
+
"max_memory_allocated (GB)": 94.39,
|
| 453 |
+
"memory_allocated (GB)": 51.7,
|
| 454 |
+
"rewards/accuracies": 0.98828125,
|
| 455 |
+
"rewards/chosen": -1.921875,
|
| 456 |
+
"rewards/margins": 13.75,
|
| 457 |
+
"rewards/rejected": -15.6875,
|
| 458 |
+
"step": 250,
|
| 459 |
+
"total_memory_available (GB)": 94.62
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.34,
|
| 463 |
+
"grad_norm": 1.0764814615249634,
|
| 464 |
+
"learning_rate": 0.00046201202403910646,
|
| 465 |
+
"logits/chosen": -3.234375,
|
| 466 |
+
"logits/rejected": -3.21875,
|
| 467 |
+
"logps/chosen": -298.0,
|
| 468 |
+
"logps/rejected": -498.0,
|
| 469 |
+
"loss": 0.0307,
|
| 470 |
+
"max_memory_allocated (GB)": 94.39,
|
| 471 |
+
"memory_allocated (GB)": 51.7,
|
| 472 |
+
"rewards/accuracies": 0.9921875,
|
| 473 |
+
"rewards/chosen": -2.421875,
|
| 474 |
+
"rewards/margins": 15.5,
|
| 475 |
+
"rewards/rejected": -17.875,
|
| 476 |
+
"step": 260,
|
| 477 |
+
"total_memory_available (GB)": 94.62
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 1.39,
|
| 481 |
+
"grad_norm": 0.04254325106739998,
|
| 482 |
+
"learning_rate": 0.00045725939313876043,
|
| 483 |
+
"logits/chosen": -3.078125,
|
| 484 |
+
"logits/rejected": -3.0625,
|
| 485 |
+
"logps/chosen": -304.0,
|
| 486 |
+
"logps/rejected": -548.0,
|
| 487 |
+
"loss": 0.0161,
|
| 488 |
+
"max_memory_allocated (GB)": 94.39,
|
| 489 |
+
"memory_allocated (GB)": 51.7,
|
| 490 |
+
"rewards/accuracies": 0.9921875,
|
| 491 |
+
"rewards/chosen": -2.703125,
|
| 492 |
+
"rewards/margins": 18.75,
|
| 493 |
+
"rewards/rejected": -21.5,
|
| 494 |
+
"step": 270,
|
| 495 |
+
"total_memory_available (GB)": 94.62
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.44,
|
| 499 |
+
"grad_norm": 0.2640644311904907,
|
| 500 |
+
"learning_rate": 0.0004522542485937369,
|
| 501 |
+
"logits/chosen": -3.078125,
|
| 502 |
+
"logits/rejected": -3.078125,
|
| 503 |
+
"logps/chosen": -294.0,
|
| 504 |
+
"logps/rejected": -552.0,
|
| 505 |
+
"loss": 0.0116,
|
| 506 |
+
"max_memory_allocated (GB)": 94.39,
|
| 507 |
+
"memory_allocated (GB)": 51.7,
|
| 508 |
+
"rewards/accuracies": 0.99609375,
|
| 509 |
+
"rewards/chosen": -2.203125,
|
| 510 |
+
"rewards/margins": 19.5,
|
| 511 |
+
"rewards/rejected": -21.75,
|
| 512 |
+
"step": 280,
|
| 513 |
+
"total_memory_available (GB)": 94.62
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.49,
|
| 517 |
+
"grad_norm": 0.6378334164619446,
|
| 518 |
+
"learning_rate": 0.00044700268840168044,
|
| 519 |
+
"logits/chosen": -3.046875,
|
| 520 |
+
"logits/rejected": -3.046875,
|
| 521 |
+
"logps/chosen": -296.0,
|
| 522 |
+
"logps/rejected": -600.0,
|
| 523 |
+
"loss": 0.0159,
|
| 524 |
+
"max_memory_allocated (GB)": 94.39,
|
| 525 |
+
"memory_allocated (GB)": 51.7,
|
| 526 |
+
"rewards/accuracies": 1.0,
|
| 527 |
+
"rewards/chosen": -3.53125,
|
| 528 |
+
"rewards/margins": 24.0,
|
| 529 |
+
"rewards/rejected": -27.625,
|
| 530 |
+
"step": 290,
|
| 531 |
+
"total_memory_available (GB)": 94.62
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.54,
|
| 535 |
+
"grad_norm": 0.9890326261520386,
|
| 536 |
+
"learning_rate": 0.0004415111107797445,
|
| 537 |
+
"logits/chosen": -3.078125,
|
| 538 |
+
"logits/rejected": -3.0625,
|
| 539 |
+
"logps/chosen": -320.0,
|
| 540 |
+
"logps/rejected": -588.0,
|
| 541 |
+
"loss": 0.0083,
|
| 542 |
+
"max_memory_allocated (GB)": 94.39,
|
| 543 |
+
"memory_allocated (GB)": 51.7,
|
| 544 |
+
"rewards/accuracies": 1.0,
|
| 545 |
+
"rewards/chosen": -3.640625,
|
| 546 |
+
"rewards/margins": 22.375,
|
| 547 |
+
"rewards/rejected": -26.0,
|
| 548 |
+
"step": 300,
|
| 549 |
+
"total_memory_available (GB)": 94.62
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.59,
|
| 553 |
+
"grad_norm": 2.6053104400634766,
|
| 554 |
+
"learning_rate": 0.00043578620636934855,
|
| 555 |
+
"logits/chosen": -3.109375,
|
| 556 |
+
"logits/rejected": -3.09375,
|
| 557 |
+
"logps/chosen": -338.0,
|
| 558 |
+
"logps/rejected": -628.0,
|
| 559 |
+
"loss": 0.0038,
|
| 560 |
+
"max_memory_allocated (GB)": 94.39,
|
| 561 |
+
"memory_allocated (GB)": 51.7,
|
| 562 |
+
"rewards/accuracies": 1.0,
|
| 563 |
+
"rewards/chosen": -5.0,
|
| 564 |
+
"rewards/margins": 24.75,
|
| 565 |
+
"rewards/rejected": -29.75,
|
| 566 |
+
"step": 310,
|
| 567 |
+
"total_memory_available (GB)": 94.62
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.64,
|
| 571 |
+
"grad_norm": 1.0414825677871704,
|
| 572 |
+
"learning_rate": 0.0004298349500846628,
|
| 573 |
+
"logits/chosen": -3.25,
|
| 574 |
+
"logits/rejected": -3.21875,
|
| 575 |
+
"logps/chosen": -302.0,
|
| 576 |
+
"logps/rejected": -540.0,
|
| 577 |
+
"loss": 0.0174,
|
| 578 |
+
"max_memory_allocated (GB)": 94.39,
|
| 579 |
+
"memory_allocated (GB)": 51.7,
|
| 580 |
+
"rewards/accuracies": 0.9921875,
|
| 581 |
+
"rewards/chosen": -3.015625,
|
| 582 |
+
"rewards/margins": 18.25,
|
| 583 |
+
"rewards/rejected": -21.25,
|
| 584 |
+
"step": 320,
|
| 585 |
+
"total_memory_available (GB)": 94.62
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.7,
|
| 589 |
+
"grad_norm": 1.3290166854858398,
|
| 590 |
+
"learning_rate": 0.00042366459261474935,
|
| 591 |
+
"logits/chosen": -3.171875,
|
| 592 |
+
"logits/rejected": -3.140625,
|
| 593 |
+
"logps/chosen": -334.0,
|
| 594 |
+
"logps/rejected": -568.0,
|
| 595 |
+
"loss": 0.0209,
|
| 596 |
+
"max_memory_allocated (GB)": 94.39,
|
| 597 |
+
"memory_allocated (GB)": 51.7,
|
| 598 |
+
"rewards/accuracies": 0.98828125,
|
| 599 |
+
"rewards/chosen": -4.1875,
|
| 600 |
+
"rewards/margins": 19.875,
|
| 601 |
+
"rewards/rejected": -24.125,
|
| 602 |
+
"step": 330,
|
| 603 |
+
"total_memory_available (GB)": 94.62
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.75,
|
| 607 |
+
"grad_norm": 3.274353504180908,
|
| 608 |
+
"learning_rate": 0.0004172826515897146,
|
| 609 |
+
"logits/chosen": -3.203125,
|
| 610 |
+
"logits/rejected": -3.171875,
|
| 611 |
+
"logps/chosen": -316.0,
|
| 612 |
+
"logps/rejected": -604.0,
|
| 613 |
+
"loss": 0.0095,
|
| 614 |
+
"max_memory_allocated (GB)": 94.39,
|
| 615 |
+
"memory_allocated (GB)": 51.7,
|
| 616 |
+
"rewards/accuracies": 1.0,
|
| 617 |
+
"rewards/chosen": -2.96875,
|
| 618 |
+
"rewards/margins": 23.625,
|
| 619 |
+
"rewards/rejected": -26.625,
|
| 620 |
+
"step": 340,
|
| 621 |
+
"total_memory_available (GB)": 94.62
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 1.8,
|
| 625 |
+
"grad_norm": 1.1814924478530884,
|
| 626 |
+
"learning_rate": 0.0004106969024216348,
|
| 627 |
+
"logits/chosen": -3.125,
|
| 628 |
+
"logits/rejected": -3.09375,
|
| 629 |
+
"logps/chosen": -294.0,
|
| 630 |
+
"logps/rejected": -612.0,
|
| 631 |
+
"loss": 0.0129,
|
| 632 |
+
"max_memory_allocated (GB)": 94.39,
|
| 633 |
+
"memory_allocated (GB)": 51.7,
|
| 634 |
+
"rewards/accuracies": 0.99609375,
|
| 635 |
+
"rewards/chosen": -3.0625,
|
| 636 |
+
"rewards/margins": 25.625,
|
| 637 |
+
"rewards/rejected": -28.625,
|
| 638 |
+
"step": 350,
|
| 639 |
+
"total_memory_available (GB)": 94.62
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.85,
|
| 643 |
+
"grad_norm": 0.10757168382406235,
|
| 644 |
+
"learning_rate": 0.00040391536883141455,
|
| 645 |
+
"logits/chosen": -3.046875,
|
| 646 |
+
"logits/rejected": -3.046875,
|
| 647 |
+
"logps/chosen": -306.0,
|
| 648 |
+
"logps/rejected": -544.0,
|
| 649 |
+
"loss": 0.0201,
|
| 650 |
+
"max_memory_allocated (GB)": 94.39,
|
| 651 |
+
"memory_allocated (GB)": 51.7,
|
| 652 |
+
"rewards/accuracies": 0.99609375,
|
| 653 |
+
"rewards/chosen": -3.078125,
|
| 654 |
+
"rewards/margins": 17.875,
|
| 655 |
+
"rewards/rejected": -21.0,
|
| 656 |
+
"step": 360,
|
| 657 |
+
"total_memory_available (GB)": 94.62
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.9,
|
| 661 |
+
"grad_norm": 0.026544878259301186,
|
| 662 |
+
"learning_rate": 0.0003969463130731183,
|
| 663 |
+
"logits/chosen": -2.953125,
|
| 664 |
+
"logits/rejected": -2.921875,
|
| 665 |
+
"logps/chosen": -304.0,
|
| 666 |
+
"logps/rejected": -548.0,
|
| 667 |
+
"loss": 0.0068,
|
| 668 |
+
"max_memory_allocated (GB)": 94.39,
|
| 669 |
+
"memory_allocated (GB)": 51.7,
|
| 670 |
+
"rewards/accuracies": 0.99609375,
|
| 671 |
+
"rewards/chosen": -3.171875,
|
| 672 |
+
"rewards/margins": 18.625,
|
| 673 |
+
"rewards/rejected": -21.75,
|
| 674 |
+
"step": 370,
|
| 675 |
+
"total_memory_available (GB)": 94.62
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.95,
|
| 679 |
+
"grad_norm": 0.012133469805121422,
|
| 680 |
+
"learning_rate": 0.0003897982258676867,
|
| 681 |
+
"logits/chosen": -2.875,
|
| 682 |
+
"logits/rejected": -2.828125,
|
| 683 |
+
"logps/chosen": -316.0,
|
| 684 |
+
"logps/rejected": -540.0,
|
| 685 |
+
"loss": 0.0172,
|
| 686 |
+
"max_memory_allocated (GB)": 94.39,
|
| 687 |
+
"memory_allocated (GB)": 51.7,
|
| 688 |
+
"rewards/accuracies": 0.99609375,
|
| 689 |
+
"rewards/chosen": -3.203125,
|
| 690 |
+
"rewards/margins": 18.0,
|
| 691 |
+
"rewards/rejected": -21.25,
|
| 692 |
+
"step": 380,
|
| 693 |
+
"total_memory_available (GB)": 94.62
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 2.0,
|
| 697 |
+
"grad_norm": 0.05422056466341019,
|
| 698 |
+
"learning_rate": 0.00038247981605830125,
|
| 699 |
+
"logits/chosen": -2.9375,
|
| 700 |
+
"logits/rejected": -2.859375,
|
| 701 |
+
"logps/chosen": -306.0,
|
| 702 |
+
"logps/rejected": -548.0,
|
| 703 |
+
"loss": 0.0067,
|
| 704 |
+
"max_memory_allocated (GB)": 94.39,
|
| 705 |
+
"memory_allocated (GB)": 51.7,
|
| 706 |
+
"rewards/accuracies": 0.99609375,
|
| 707 |
+
"rewards/chosen": -3.25,
|
| 708 |
+
"rewards/margins": 19.875,
|
| 709 |
+
"rewards/rejected": -23.125,
|
| 710 |
+
"step": 390,
|
| 711 |
+
"total_memory_available (GB)": 94.62
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.06,
|
| 715 |
+
"grad_norm": 0.5953335762023926,
|
| 716 |
+
"learning_rate": 0.000375,
|
| 717 |
+
"logits/chosen": -3.0625,
|
| 718 |
+
"logits/rejected": -3.015625,
|
| 719 |
+
"logps/chosen": -300.0,
|
| 720 |
+
"logps/rejected": -588.0,
|
| 721 |
+
"loss": 0.0196,
|
| 722 |
+
"max_memory_allocated (GB)": 94.39,
|
| 723 |
+
"memory_allocated (GB)": 51.7,
|
| 724 |
+
"rewards/accuracies": 0.99609375,
|
| 725 |
+
"rewards/chosen": -3.6875,
|
| 726 |
+
"rewards/margins": 21.625,
|
| 727 |
+
"rewards/rejected": -25.25,
|
| 728 |
+
"step": 400,
|
| 729 |
+
"total_memory_available (GB)": 94.62
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.11,
|
| 733 |
+
"grad_norm": 4.781039714813232,
|
| 734 |
+
"learning_rate": 0.0003673678906964727,
|
| 735 |
+
"logits/chosen": -3.015625,
|
| 736 |
+
"logits/rejected": -3.0,
|
| 737 |
+
"logps/chosen": -300.0,
|
| 738 |
+
"logps/rejected": -564.0,
|
| 739 |
+
"loss": 0.0067,
|
| 740 |
+
"max_memory_allocated (GB)": 94.39,
|
| 741 |
+
"memory_allocated (GB)": 51.7,
|
| 742 |
+
"rewards/accuracies": 0.99609375,
|
| 743 |
+
"rewards/chosen": -3.03125,
|
| 744 |
+
"rewards/margins": 19.25,
|
| 745 |
+
"rewards/rejected": -22.375,
|
| 746 |
+
"step": 410,
|
| 747 |
+
"total_memory_available (GB)": 94.62
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.16,
|
| 751 |
+
"grad_norm": 0.0018408839823678136,
|
| 752 |
+
"learning_rate": 0.00035959278669726934,
|
| 753 |
+
"logits/chosen": -2.953125,
|
| 754 |
+
"logits/rejected": -2.921875,
|
| 755 |
+
"logps/chosen": -300.0,
|
| 756 |
+
"logps/rejected": -600.0,
|
| 757 |
+
"loss": 0.0012,
|
| 758 |
+
"max_memory_allocated (GB)": 94.39,
|
| 759 |
+
"memory_allocated (GB)": 51.7,
|
| 760 |
+
"rewards/accuracies": 1.0,
|
| 761 |
+
"rewards/chosen": -3.5625,
|
| 762 |
+
"rewards/margins": 22.75,
|
| 763 |
+
"rewards/rejected": -26.25,
|
| 764 |
+
"step": 420,
|
| 765 |
+
"total_memory_available (GB)": 94.62
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.21,
|
| 769 |
+
"grad_norm": 0.0082088652998209,
|
| 770 |
+
"learning_rate": 0.0003516841607689501,
|
| 771 |
+
"logits/chosen": -3.0625,
|
| 772 |
+
"logits/rejected": -3.015625,
|
| 773 |
+
"logps/chosen": -308.0,
|
| 774 |
+
"logps/rejected": -560.0,
|
| 775 |
+
"loss": 0.0058,
|
| 776 |
+
"max_memory_allocated (GB)": 94.39,
|
| 777 |
+
"memory_allocated (GB)": 51.7,
|
| 778 |
+
"rewards/accuracies": 0.99609375,
|
| 779 |
+
"rewards/chosen": -2.828125,
|
| 780 |
+
"rewards/margins": 21.0,
|
| 781 |
+
"rewards/rejected": -23.875,
|
| 782 |
+
"step": 430,
|
| 783 |
+
"total_memory_available (GB)": 94.62
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.26,
|
| 787 |
+
"grad_norm": 0.0047104801051318645,
|
| 788 |
+
"learning_rate": 0.00034365164835397803,
|
| 789 |
+
"logits/chosen": -3.15625,
|
| 790 |
+
"logits/rejected": -3.140625,
|
| 791 |
+
"logps/chosen": -290.0,
|
| 792 |
+
"logps/rejected": -536.0,
|
| 793 |
+
"loss": 0.0025,
|
| 794 |
+
"max_memory_allocated (GB)": 94.39,
|
| 795 |
+
"memory_allocated (GB)": 51.7,
|
| 796 |
+
"rewards/accuracies": 1.0,
|
| 797 |
+
"rewards/chosen": -2.171875,
|
| 798 |
+
"rewards/margins": 18.375,
|
| 799 |
+
"rewards/rejected": -20.5,
|
| 800 |
+
"step": 440,
|
| 801 |
+
"total_memory_available (GB)": 94.62
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.31,
|
| 805 |
+
"grad_norm": 0.10709760338068008,
|
| 806 |
+
"learning_rate": 0.0003355050358314172,
|
| 807 |
+
"logits/chosen": -3.171875,
|
| 808 |
+
"logits/rejected": -3.140625,
|
| 809 |
+
"logps/chosen": -310.0,
|
| 810 |
+
"logps/rejected": -560.0,
|
| 811 |
+
"loss": 0.0015,
|
| 812 |
+
"max_memory_allocated (GB)": 94.39,
|
| 813 |
+
"memory_allocated (GB)": 51.7,
|
| 814 |
+
"rewards/accuracies": 1.0,
|
| 815 |
+
"rewards/chosen": -2.75,
|
| 816 |
+
"rewards/margins": 21.25,
|
| 817 |
+
"rewards/rejected": -24.0,
|
| 818 |
+
"step": 450,
|
| 819 |
+
"total_memory_available (GB)": 94.62
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.36,
|
| 823 |
+
"grad_norm": 0.016811428591609,
|
| 824 |
+
"learning_rate": 0.00032725424859373687,
|
| 825 |
+
"logits/chosen": -3.125,
|
| 826 |
+
"logits/rejected": -3.078125,
|
| 827 |
+
"logps/chosen": -304.0,
|
| 828 |
+
"logps/rejected": -596.0,
|
| 829 |
+
"loss": 0.0003,
|
| 830 |
+
"max_memory_allocated (GB)": 94.39,
|
| 831 |
+
"memory_allocated (GB)": 51.7,
|
| 832 |
+
"rewards/accuracies": 1.0,
|
| 833 |
+
"rewards/chosen": -3.40625,
|
| 834 |
+
"rewards/margins": 22.875,
|
| 835 |
+
"rewards/rejected": -26.25,
|
| 836 |
+
"step": 460,
|
| 837 |
+
"total_memory_available (GB)": 94.62
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.41,
|
| 841 |
+
"grad_norm": 0.05604538321495056,
|
| 842 |
+
"learning_rate": 0.0003189093389542498,
|
| 843 |
+
"logits/chosen": -3.078125,
|
| 844 |
+
"logits/rejected": -3.03125,
|
| 845 |
+
"logps/chosen": -318.0,
|
| 846 |
+
"logps/rejected": -612.0,
|
| 847 |
+
"loss": 0.0003,
|
| 848 |
+
"max_memory_allocated (GB)": 94.39,
|
| 849 |
+
"memory_allocated (GB)": 51.7,
|
| 850 |
+
"rewards/accuracies": 1.0,
|
| 851 |
+
"rewards/chosen": -3.828125,
|
| 852 |
+
"rewards/margins": 24.25,
|
| 853 |
+
"rewards/rejected": -28.0,
|
| 854 |
+
"step": 470,
|
| 855 |
+
"total_memory_available (GB)": 94.62
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.47,
|
| 859 |
+
"grad_norm": 0.10262610763311386,
|
| 860 |
+
"learning_rate": 0.0003104804738999169,
|
| 861 |
+
"logits/chosen": -3.0,
|
| 862 |
+
"logits/rejected": -2.9375,
|
| 863 |
+
"logps/chosen": -304.0,
|
| 864 |
+
"logps/rejected": -608.0,
|
| 865 |
+
"loss": 0.0063,
|
| 866 |
+
"max_memory_allocated (GB)": 94.39,
|
| 867 |
+
"memory_allocated (GB)": 51.7,
|
| 868 |
+
"rewards/accuracies": 0.99609375,
|
| 869 |
+
"rewards/chosen": -4.0,
|
| 870 |
+
"rewards/margins": 23.125,
|
| 871 |
+
"rewards/rejected": -27.125,
|
| 872 |
+
"step": 480,
|
| 873 |
+
"total_memory_available (GB)": 94.62
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.52,
|
| 877 |
+
"grad_norm": 2.782456874847412,
|
| 878 |
+
"learning_rate": 0.0003019779227044398,
|
| 879 |
+
"logits/chosen": -2.859375,
|
| 880 |
+
"logits/rejected": -2.75,
|
| 881 |
+
"logps/chosen": -322.0,
|
| 882 |
+
"logps/rejected": -544.0,
|
| 883 |
+
"loss": 0.0095,
|
| 884 |
+
"max_memory_allocated (GB)": 94.39,
|
| 885 |
+
"memory_allocated (GB)": 51.7,
|
| 886 |
+
"rewards/accuracies": 0.99609375,
|
| 887 |
+
"rewards/chosen": -3.203125,
|
| 888 |
+
"rewards/margins": 18.875,
|
| 889 |
+
"rewards/rejected": -22.125,
|
| 890 |
+
"step": 490,
|
| 891 |
+
"total_memory_available (GB)": 94.62
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.57,
|
| 895 |
+
"grad_norm": 0.1368335485458374,
|
| 896 |
+
"learning_rate": 0.00029341204441673266,
|
| 897 |
+
"logits/chosen": -3.03125,
|
| 898 |
+
"logits/rejected": -2.9375,
|
| 899 |
+
"logps/chosen": -296.0,
|
| 900 |
+
"logps/rejected": -556.0,
|
| 901 |
+
"loss": 0.0029,
|
| 902 |
+
"max_memory_allocated (GB)": 94.39,
|
| 903 |
+
"memory_allocated (GB)": 51.7,
|
| 904 |
+
"rewards/accuracies": 1.0,
|
| 905 |
+
"rewards/chosen": -2.359375,
|
| 906 |
+
"rewards/margins": 19.875,
|
| 907 |
+
"rewards/rejected": -22.125,
|
| 908 |
+
"step": 500,
|
| 909 |
+
"total_memory_available (GB)": 94.62
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.62,
|
| 913 |
+
"grad_norm": 0.009250438772141933,
|
| 914 |
+
"learning_rate": 0.00028479327524001636,
|
| 915 |
+
"logits/chosen": -3.171875,
|
| 916 |
+
"logits/rejected": -3.109375,
|
| 917 |
+
"logps/chosen": -318.0,
|
| 918 |
+
"logps/rejected": -552.0,
|
| 919 |
+
"loss": 0.0044,
|
| 920 |
+
"max_memory_allocated (GB)": 94.39,
|
| 921 |
+
"memory_allocated (GB)": 51.7,
|
| 922 |
+
"rewards/accuracies": 1.0,
|
| 923 |
+
"rewards/chosen": -2.796875,
|
| 924 |
+
"rewards/margins": 19.75,
|
| 925 |
+
"rewards/rejected": -22.625,
|
| 926 |
+
"step": 510,
|
| 927 |
+
"total_memory_available (GB)": 94.62
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.67,
|
| 931 |
+
"grad_norm": 0.1493176519870758,
|
| 932 |
+
"learning_rate": 0.0002761321158169134,
|
| 933 |
+
"logits/chosen": -3.21875,
|
| 934 |
+
"logits/rejected": -3.1875,
|
| 935 |
+
"logps/chosen": -298.0,
|
| 936 |
+
"logps/rejected": -524.0,
|
| 937 |
+
"loss": 0.0048,
|
| 938 |
+
"max_memory_allocated (GB)": 94.39,
|
| 939 |
+
"memory_allocated (GB)": 51.7,
|
| 940 |
+
"rewards/accuracies": 1.0,
|
| 941 |
+
"rewards/chosen": -2.46875,
|
| 942 |
+
"rewards/margins": 17.5,
|
| 943 |
+
"rewards/rejected": -20.0,
|
| 944 |
+
"step": 520,
|
| 945 |
+
"total_memory_available (GB)": 94.62
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 2.72,
|
| 949 |
+
"grad_norm": 0.06779959052801132,
|
| 950 |
+
"learning_rate": 0.0002674391184360313,
|
| 951 |
+
"logits/chosen": -3.234375,
|
| 952 |
+
"logits/rejected": -3.21875,
|
| 953 |
+
"logps/chosen": -298.0,
|
| 954 |
+
"logps/rejected": -528.0,
|
| 955 |
+
"loss": 0.0038,
|
| 956 |
+
"max_memory_allocated (GB)": 94.39,
|
| 957 |
+
"memory_allocated (GB)": 51.7,
|
| 958 |
+
"rewards/accuracies": 1.0,
|
| 959 |
+
"rewards/chosen": -2.4375,
|
| 960 |
+
"rewards/margins": 17.625,
|
| 961 |
+
"rewards/rejected": -20.125,
|
| 962 |
+
"step": 530,
|
| 963 |
+
"total_memory_available (GB)": 94.62
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 2.77,
|
| 967 |
+
"grad_norm": 0.3520822525024414,
|
| 968 |
+
"learning_rate": 0.0002587248741756253,
|
| 969 |
+
"logits/chosen": -3.25,
|
| 970 |
+
"logits/rejected": -3.1875,
|
| 971 |
+
"logps/chosen": -302.0,
|
| 972 |
+
"logps/rejected": -536.0,
|
| 973 |
+
"loss": 0.0021,
|
| 974 |
+
"max_memory_allocated (GB)": 94.39,
|
| 975 |
+
"memory_allocated (GB)": 51.7,
|
| 976 |
+
"rewards/accuracies": 1.0,
|
| 977 |
+
"rewards/chosen": -2.734375,
|
| 978 |
+
"rewards/margins": 19.5,
|
| 979 |
+
"rewards/rejected": -22.25,
|
| 980 |
+
"step": 540,
|
| 981 |
+
"total_memory_available (GB)": 94.62
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 2.83,
|
| 985 |
+
"grad_norm": 0.020363105461001396,
|
| 986 |
+
"learning_rate": 0.00025,
|
| 987 |
+
"logits/chosen": -3.15625,
|
| 988 |
+
"logits/rejected": -3.125,
|
| 989 |
+
"logps/chosen": -310.0,
|
| 990 |
+
"logps/rejected": -572.0,
|
| 991 |
+
"loss": 0.0016,
|
| 992 |
+
"max_memory_allocated (GB)": 94.39,
|
| 993 |
+
"memory_allocated (GB)": 51.7,
|
| 994 |
+
"rewards/accuracies": 1.0,
|
| 995 |
+
"rewards/chosen": -3.40625,
|
| 996 |
+
"rewards/margins": 21.625,
|
| 997 |
+
"rewards/rejected": -25.0,
|
| 998 |
+
"step": 550,
|
| 999 |
+
"total_memory_available (GB)": 94.62
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 2.88,
|
| 1003 |
+
"grad_norm": 0.01094722468405962,
|
| 1004 |
+
"learning_rate": 0.00024127512582437484,
|
| 1005 |
+
"logits/chosen": -3.109375,
|
| 1006 |
+
"logits/rejected": -3.0625,
|
| 1007 |
+
"logps/chosen": -316.0,
|
| 1008 |
+
"logps/rejected": -640.0,
|
| 1009 |
+
"loss": 0.0016,
|
| 1010 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1011 |
+
"memory_allocated (GB)": 51.7,
|
| 1012 |
+
"rewards/accuracies": 1.0,
|
| 1013 |
+
"rewards/chosen": -3.90625,
|
| 1014 |
+
"rewards/margins": 26.125,
|
| 1015 |
+
"rewards/rejected": -30.0,
|
| 1016 |
+
"step": 560,
|
| 1017 |
+
"total_memory_available (GB)": 94.62
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.93,
|
| 1021 |
+
"grad_norm": 0.10121448338031769,
|
| 1022 |
+
"learning_rate": 0.00023256088156396867,
|
| 1023 |
+
"logits/chosen": -3.078125,
|
| 1024 |
+
"logits/rejected": -3.03125,
|
| 1025 |
+
"logps/chosen": -332.0,
|
| 1026 |
+
"logps/rejected": -628.0,
|
| 1027 |
+
"loss": 0.0016,
|
| 1028 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1029 |
+
"memory_allocated (GB)": 51.7,
|
| 1030 |
+
"rewards/accuracies": 1.0,
|
| 1031 |
+
"rewards/chosen": -3.875,
|
| 1032 |
+
"rewards/margins": 26.375,
|
| 1033 |
+
"rewards/rejected": -30.25,
|
| 1034 |
+
"step": 570,
|
| 1035 |
+
"total_memory_available (GB)": 94.62
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 2.98,
|
| 1039 |
+
"grad_norm": 0.01434832252562046,
|
| 1040 |
+
"learning_rate": 0.00022386788418308668,
|
| 1041 |
+
"logits/chosen": -3.0,
|
| 1042 |
+
"logits/rejected": -2.90625,
|
| 1043 |
+
"logps/chosen": -328.0,
|
| 1044 |
+
"logps/rejected": -640.0,
|
| 1045 |
+
"loss": 0.0004,
|
| 1046 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1047 |
+
"memory_allocated (GB)": 51.7,
|
| 1048 |
+
"rewards/accuracies": 1.0,
|
| 1049 |
+
"rewards/chosen": -4.375,
|
| 1050 |
+
"rewards/margins": 26.75,
|
| 1051 |
+
"rewards/rejected": -31.125,
|
| 1052 |
+
"step": 580,
|
| 1053 |
+
"total_memory_available (GB)": 94.62
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 3.03,
|
| 1057 |
+
"grad_norm": 0.0004497322079259902,
|
| 1058 |
+
"learning_rate": 0.0002152067247599837,
|
| 1059 |
+
"logits/chosen": -2.953125,
|
| 1060 |
+
"logits/rejected": -2.875,
|
| 1061 |
+
"logps/chosen": -304.0,
|
| 1062 |
+
"logps/rejected": -656.0,
|
| 1063 |
+
"loss": 0.0005,
|
| 1064 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1065 |
+
"memory_allocated (GB)": 51.7,
|
| 1066 |
+
"rewards/accuracies": 1.0,
|
| 1067 |
+
"rewards/chosen": -4.375,
|
| 1068 |
+
"rewards/margins": 28.25,
|
| 1069 |
+
"rewards/rejected": -32.75,
|
| 1070 |
+
"step": 590,
|
| 1071 |
+
"total_memory_available (GB)": 94.62
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 3.08,
|
| 1075 |
+
"grad_norm": 0.010737881064414978,
|
| 1076 |
+
"learning_rate": 0.00020658795558326743,
|
| 1077 |
+
"logits/chosen": -2.984375,
|
| 1078 |
+
"logits/rejected": -2.859375,
|
| 1079 |
+
"logps/chosen": -330.0,
|
| 1080 |
+
"logps/rejected": -676.0,
|
| 1081 |
+
"loss": 0.0008,
|
| 1082 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1083 |
+
"memory_allocated (GB)": 51.7,
|
| 1084 |
+
"rewards/accuracies": 1.0,
|
| 1085 |
+
"rewards/chosen": -5.03125,
|
| 1086 |
+
"rewards/margins": 29.375,
|
| 1087 |
+
"rewards/rejected": -34.5,
|
| 1088 |
+
"step": 600,
|
| 1089 |
+
"total_memory_available (GB)": 94.62
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 3.13,
|
| 1093 |
+
"grad_norm": 0.3118731379508972,
|
| 1094 |
+
"learning_rate": 0.0001980220772955602,
|
| 1095 |
+
"logits/chosen": -3.015625,
|
| 1096 |
+
"logits/rejected": -2.9375,
|
| 1097 |
+
"logps/chosen": -310.0,
|
| 1098 |
+
"logps/rejected": -648.0,
|
| 1099 |
+
"loss": 0.0011,
|
| 1100 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1101 |
+
"memory_allocated (GB)": 51.7,
|
| 1102 |
+
"rewards/accuracies": 1.0,
|
| 1103 |
+
"rewards/chosen": -3.828125,
|
| 1104 |
+
"rewards/margins": 27.75,
|
| 1105 |
+
"rewards/rejected": -31.5,
|
| 1106 |
+
"step": 610,
|
| 1107 |
+
"total_memory_available (GB)": 94.62
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.19,
|
| 1111 |
+
"grad_norm": 0.04152989387512207,
|
| 1112 |
+
"learning_rate": 0.0001895195261000831,
|
| 1113 |
+
"logits/chosen": -3.09375,
|
| 1114 |
+
"logits/rejected": -3.03125,
|
| 1115 |
+
"logps/chosen": -302.0,
|
| 1116 |
+
"logps/rejected": -604.0,
|
| 1117 |
+
"loss": 0.0002,
|
| 1118 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1119 |
+
"memory_allocated (GB)": 51.7,
|
| 1120 |
+
"rewards/accuracies": 1.0,
|
| 1121 |
+
"rewards/chosen": -2.75,
|
| 1122 |
+
"rewards/margins": 24.75,
|
| 1123 |
+
"rewards/rejected": -27.5,
|
| 1124 |
+
"step": 620,
|
| 1125 |
+
"total_memory_available (GB)": 94.62
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 3.24,
|
| 1129 |
+
"grad_norm": 0.00012200848141219467,
|
| 1130 |
+
"learning_rate": 0.00018109066104575022,
|
| 1131 |
+
"logits/chosen": -3.125,
|
| 1132 |
+
"logits/rejected": -3.0625,
|
| 1133 |
+
"logps/chosen": -324.0,
|
| 1134 |
+
"logps/rejected": -584.0,
|
| 1135 |
+
"loss": 0.0013,
|
| 1136 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1137 |
+
"memory_allocated (GB)": 51.7,
|
| 1138 |
+
"rewards/accuracies": 1.0,
|
| 1139 |
+
"rewards/chosen": -2.875,
|
| 1140 |
+
"rewards/margins": 23.375,
|
| 1141 |
+
"rewards/rejected": -26.25,
|
| 1142 |
+
"step": 630,
|
| 1143 |
+
"total_memory_available (GB)": 94.62
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 3.29,
|
| 1147 |
+
"grad_norm": 0.037591688334941864,
|
| 1148 |
+
"learning_rate": 0.00017274575140626317,
|
| 1149 |
+
"logits/chosen": -3.09375,
|
| 1150 |
+
"logits/rejected": -3.046875,
|
| 1151 |
+
"logps/chosen": -308.0,
|
| 1152 |
+
"logps/rejected": -608.0,
|
| 1153 |
+
"loss": 0.0001,
|
| 1154 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1155 |
+
"memory_allocated (GB)": 51.7,
|
| 1156 |
+
"rewards/accuracies": 1.0,
|
| 1157 |
+
"rewards/chosen": -2.9375,
|
| 1158 |
+
"rewards/margins": 24.375,
|
| 1159 |
+
"rewards/rejected": -27.25,
|
| 1160 |
+
"step": 640,
|
| 1161 |
+
"total_memory_available (GB)": 94.62
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 3.34,
|
| 1165 |
+
"grad_norm": 0.0018848556792363524,
|
| 1166 |
+
"learning_rate": 0.00016449496416858284,
|
| 1167 |
+
"logits/chosen": -3.109375,
|
| 1168 |
+
"logits/rejected": -3.03125,
|
| 1169 |
+
"logps/chosen": -308.0,
|
| 1170 |
+
"logps/rejected": -600.0,
|
| 1171 |
+
"loss": 0.0002,
|
| 1172 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1173 |
+
"memory_allocated (GB)": 51.7,
|
| 1174 |
+
"rewards/accuracies": 1.0,
|
| 1175 |
+
"rewards/chosen": -3.171875,
|
| 1176 |
+
"rewards/margins": 24.25,
|
| 1177 |
+
"rewards/rejected": -27.5,
|
| 1178 |
+
"step": 650,
|
| 1179 |
+
"total_memory_available (GB)": 94.62
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 3.39,
|
| 1183 |
+
"grad_norm": 0.0032229481730610132,
|
| 1184 |
+
"learning_rate": 0.00015634835164602198,
|
| 1185 |
+
"logits/chosen": -3.078125,
|
| 1186 |
+
"logits/rejected": -3.046875,
|
| 1187 |
+
"logps/chosen": -312.0,
|
| 1188 |
+
"logps/rejected": -612.0,
|
| 1189 |
+
"loss": 0.0002,
|
| 1190 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1191 |
+
"memory_allocated (GB)": 51.7,
|
| 1192 |
+
"rewards/accuracies": 1.0,
|
| 1193 |
+
"rewards/chosen": -3.203125,
|
| 1194 |
+
"rewards/margins": 24.75,
|
| 1195 |
+
"rewards/rejected": -28.0,
|
| 1196 |
+
"step": 660,
|
| 1197 |
+
"total_memory_available (GB)": 94.62
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 3.44,
|
| 1201 |
+
"grad_norm": 0.0005074635264463723,
|
| 1202 |
+
"learning_rate": 0.00014831583923105,
|
| 1203 |
+
"logits/chosen": -3.09375,
|
| 1204 |
+
"logits/rejected": -3.015625,
|
| 1205 |
+
"logps/chosen": -316.0,
|
| 1206 |
+
"logps/rejected": -596.0,
|
| 1207 |
+
"loss": 0.0,
|
| 1208 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1209 |
+
"memory_allocated (GB)": 51.7,
|
| 1210 |
+
"rewards/accuracies": 1.0,
|
| 1211 |
+
"rewards/chosen": -3.203125,
|
| 1212 |
+
"rewards/margins": 24.625,
|
| 1213 |
+
"rewards/rejected": -27.875,
|
| 1214 |
+
"step": 670,
|
| 1215 |
+
"total_memory_available (GB)": 94.62
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 3.49,
|
| 1219 |
+
"grad_norm": 0.007239128462970257,
|
| 1220 |
+
"learning_rate": 0.00014040721330273062,
|
| 1221 |
+
"logits/chosen": -3.109375,
|
| 1222 |
+
"logits/rejected": -3.046875,
|
| 1223 |
+
"logps/chosen": -310.0,
|
| 1224 |
+
"logps/rejected": -604.0,
|
| 1225 |
+
"loss": 0.0001,
|
| 1226 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1227 |
+
"memory_allocated (GB)": 51.7,
|
| 1228 |
+
"rewards/accuracies": 1.0,
|
| 1229 |
+
"rewards/chosen": -3.265625,
|
| 1230 |
+
"rewards/margins": 24.75,
|
| 1231 |
+
"rewards/rejected": -28.0,
|
| 1232 |
+
"step": 680,
|
| 1233 |
+
"total_memory_available (GB)": 94.62
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 3.55,
|
| 1237 |
+
"grad_norm": 0.0034740234259516,
|
| 1238 |
+
"learning_rate": 0.00013263210930352737,
|
| 1239 |
+
"logits/chosen": -3.078125,
|
| 1240 |
+
"logits/rejected": -3.03125,
|
| 1241 |
+
"logps/chosen": -300.0,
|
| 1242 |
+
"logps/rejected": -624.0,
|
| 1243 |
+
"loss": 0.0002,
|
| 1244 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1245 |
+
"memory_allocated (GB)": 51.7,
|
| 1246 |
+
"rewards/accuracies": 1.0,
|
| 1247 |
+
"rewards/chosen": -3.15625,
|
| 1248 |
+
"rewards/margins": 25.375,
|
| 1249 |
+
"rewards/rejected": -28.5,
|
| 1250 |
+
"step": 690,
|
| 1251 |
+
"total_memory_available (GB)": 94.62
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 3.6,
|
| 1255 |
+
"grad_norm": 0.0001397097803419456,
|
| 1256 |
+
"learning_rate": 0.00012500000000000006,
|
| 1257 |
+
"logits/chosen": -3.078125,
|
| 1258 |
+
"logits/rejected": -3.046875,
|
| 1259 |
+
"logps/chosen": -304.0,
|
| 1260 |
+
"logps/rejected": -620.0,
|
| 1261 |
+
"loss": 0.0001,
|
| 1262 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1263 |
+
"memory_allocated (GB)": 51.7,
|
| 1264 |
+
"rewards/accuracies": 1.0,
|
| 1265 |
+
"rewards/chosen": -3.1875,
|
| 1266 |
+
"rewards/margins": 24.875,
|
| 1267 |
+
"rewards/rejected": -28.125,
|
| 1268 |
+
"step": 700,
|
| 1269 |
+
"total_memory_available (GB)": 94.62
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 3.65,
|
| 1273 |
+
"grad_norm": 0.0016049661207944155,
|
| 1274 |
+
"learning_rate": 0.0001175201839416988,
|
| 1275 |
+
"logits/chosen": -3.09375,
|
| 1276 |
+
"logits/rejected": -3.03125,
|
| 1277 |
+
"logps/chosen": -324.0,
|
| 1278 |
+
"logps/rejected": -608.0,
|
| 1279 |
+
"loss": 0.0001,
|
| 1280 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1281 |
+
"memory_allocated (GB)": 51.7,
|
| 1282 |
+
"rewards/accuracies": 1.0,
|
| 1283 |
+
"rewards/chosen": -3.28125,
|
| 1284 |
+
"rewards/margins": 24.75,
|
| 1285 |
+
"rewards/rejected": -28.0,
|
| 1286 |
+
"step": 710,
|
| 1287 |
+
"total_memory_available (GB)": 94.62
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 3.7,
|
| 1291 |
+
"grad_norm": 0.006374781485646963,
|
| 1292 |
+
"learning_rate": 0.00011020177413231333,
|
| 1293 |
+
"logits/chosen": -3.109375,
|
| 1294 |
+
"logits/rejected": -3.03125,
|
| 1295 |
+
"logps/chosen": -306.0,
|
| 1296 |
+
"logps/rejected": -608.0,
|
| 1297 |
+
"loss": 0.0002,
|
| 1298 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1299 |
+
"memory_allocated (GB)": 51.7,
|
| 1300 |
+
"rewards/accuracies": 1.0,
|
| 1301 |
+
"rewards/chosen": -3.28125,
|
| 1302 |
+
"rewards/margins": 25.125,
|
| 1303 |
+
"rewards/rejected": -28.375,
|
| 1304 |
+
"step": 720,
|
| 1305 |
+
"total_memory_available (GB)": 94.62
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 3.75,
|
| 1309 |
+
"grad_norm": 0.0068184914998710155,
|
| 1310 |
+
"learning_rate": 0.00010305368692688174,
|
| 1311 |
+
"logits/chosen": -3.09375,
|
| 1312 |
+
"logits/rejected": -3.046875,
|
| 1313 |
+
"logps/chosen": -296.0,
|
| 1314 |
+
"logps/rejected": -616.0,
|
| 1315 |
+
"loss": 0.0004,
|
| 1316 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1317 |
+
"memory_allocated (GB)": 51.7,
|
| 1318 |
+
"rewards/accuracies": 1.0,
|
| 1319 |
+
"rewards/chosen": -3.15625,
|
| 1320 |
+
"rewards/margins": 25.125,
|
| 1321 |
+
"rewards/rejected": -28.25,
|
| 1322 |
+
"step": 730,
|
| 1323 |
+
"total_memory_available (GB)": 94.62
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 3.8,
|
| 1327 |
+
"grad_norm": 0.008304049260914326,
|
| 1328 |
+
"learning_rate": 9.608463116858542e-05,
|
| 1329 |
+
"logits/chosen": -3.09375,
|
| 1330 |
+
"logits/rejected": -3.046875,
|
| 1331 |
+
"logps/chosen": -300.0,
|
| 1332 |
+
"logps/rejected": -592.0,
|
| 1333 |
+
"loss": 0.0002,
|
| 1334 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1335 |
+
"memory_allocated (GB)": 51.7,
|
| 1336 |
+
"rewards/accuracies": 1.0,
|
| 1337 |
+
"rewards/chosen": -3.046875,
|
| 1338 |
+
"rewards/margins": 24.0,
|
| 1339 |
+
"rewards/rejected": -27.0,
|
| 1340 |
+
"step": 740,
|
| 1341 |
+
"total_memory_available (GB)": 94.62
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 3.85,
|
| 1345 |
+
"grad_norm": 0.03573792800307274,
|
| 1346 |
+
"learning_rate": 8.930309757836516e-05,
|
| 1347 |
+
"logits/chosen": -3.09375,
|
| 1348 |
+
"logits/rejected": -3.03125,
|
| 1349 |
+
"logps/chosen": -296.0,
|
| 1350 |
+
"logps/rejected": -604.0,
|
| 1351 |
+
"loss": 0.0001,
|
| 1352 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1353 |
+
"memory_allocated (GB)": 51.7,
|
| 1354 |
+
"rewards/accuracies": 1.0,
|
| 1355 |
+
"rewards/chosen": -3.015625,
|
| 1356 |
+
"rewards/margins": 24.75,
|
| 1357 |
+
"rewards/rejected": -27.75,
|
| 1358 |
+
"step": 750,
|
| 1359 |
+
"total_memory_available (GB)": 94.62
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 3.9,
|
| 1363 |
+
"grad_norm": 0.0008675418794155121,
|
| 1364 |
+
"learning_rate": 8.271734841028553e-05,
|
| 1365 |
+
"logits/chosen": -3.09375,
|
| 1366 |
+
"logits/rejected": -3.03125,
|
| 1367 |
+
"logps/chosen": -296.0,
|
| 1368 |
+
"logps/rejected": -612.0,
|
| 1369 |
+
"loss": 0.0001,
|
| 1370 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1371 |
+
"memory_allocated (GB)": 51.7,
|
| 1372 |
+
"rewards/accuracies": 1.0,
|
| 1373 |
+
"rewards/chosen": -2.984375,
|
| 1374 |
+
"rewards/margins": 25.125,
|
| 1375 |
+
"rewards/rejected": -28.0,
|
| 1376 |
+
"step": 760,
|
| 1377 |
+
"total_memory_available (GB)": 94.62
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 3.96,
|
| 1381 |
+
"grad_norm": 4.6364646550500765e-05,
|
| 1382 |
+
"learning_rate": 7.633540738525066e-05,
|
| 1383 |
+
"logits/chosen": -3.109375,
|
| 1384 |
+
"logits/rejected": -3.046875,
|
| 1385 |
+
"logps/chosen": -320.0,
|
| 1386 |
+
"logps/rejected": -616.0,
|
| 1387 |
+
"loss": 0.0001,
|
| 1388 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1389 |
+
"memory_allocated (GB)": 51.7,
|
| 1390 |
+
"rewards/accuracies": 1.0,
|
| 1391 |
+
"rewards/chosen": -3.25,
|
| 1392 |
+
"rewards/margins": 24.875,
|
| 1393 |
+
"rewards/rejected": -28.125,
|
| 1394 |
+
"step": 770,
|
| 1395 |
+
"total_memory_available (GB)": 94.62
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.01,
|
| 1399 |
+
"grad_norm": 0.0028735387604683638,
|
| 1400 |
+
"learning_rate": 7.016504991533726e-05,
|
| 1401 |
+
"logits/chosen": -3.09375,
|
| 1402 |
+
"logits/rejected": -3.046875,
|
| 1403 |
+
"logps/chosen": -304.0,
|
| 1404 |
+
"logps/rejected": -616.0,
|
| 1405 |
+
"loss": 0.0,
|
| 1406 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1407 |
+
"memory_allocated (GB)": 51.7,
|
| 1408 |
+
"rewards/accuracies": 1.0,
|
| 1409 |
+
"rewards/chosen": -2.921875,
|
| 1410 |
+
"rewards/margins": 25.25,
|
| 1411 |
+
"rewards/rejected": -28.125,
|
| 1412 |
+
"step": 780,
|
| 1413 |
+
"total_memory_available (GB)": 94.62
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 4.06,
|
| 1417 |
+
"grad_norm": 0.00028936524176970124,
|
| 1418 |
+
"learning_rate": 6.421379363065141e-05,
|
| 1419 |
+
"logits/chosen": -3.09375,
|
| 1420 |
+
"logits/rejected": -3.046875,
|
| 1421 |
+
"logps/chosen": -302.0,
|
| 1422 |
+
"logps/rejected": -596.0,
|
| 1423 |
+
"loss": 0.0001,
|
| 1424 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1425 |
+
"memory_allocated (GB)": 51.7,
|
| 1426 |
+
"rewards/accuracies": 1.0,
|
| 1427 |
+
"rewards/chosen": -2.875,
|
| 1428 |
+
"rewards/margins": 25.0,
|
| 1429 |
+
"rewards/rejected": -27.875,
|
| 1430 |
+
"step": 790,
|
| 1431 |
+
"total_memory_available (GB)": 94.62
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.11,
|
| 1435 |
+
"grad_norm": 4.6098506572889164e-05,
|
| 1436 |
+
"learning_rate": 5.848888922025553e-05,
|
| 1437 |
+
"logits/chosen": -3.09375,
|
| 1438 |
+
"logits/rejected": -3.046875,
|
| 1439 |
+
"logps/chosen": -308.0,
|
| 1440 |
+
"logps/rejected": -616.0,
|
| 1441 |
+
"loss": 0.0,
|
| 1442 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1443 |
+
"memory_allocated (GB)": 51.7,
|
| 1444 |
+
"rewards/accuracies": 1.0,
|
| 1445 |
+
"rewards/chosen": -3.125,
|
| 1446 |
+
"rewards/margins": 25.5,
|
| 1447 |
+
"rewards/rejected": -28.625,
|
| 1448 |
+
"step": 800,
|
| 1449 |
+
"total_memory_available (GB)": 94.62
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 4.16,
|
| 1453 |
+
"grad_norm": 0.0014158826088532805,
|
| 1454 |
+
"learning_rate": 5.299731159831953e-05,
|
| 1455 |
+
"logits/chosen": -3.09375,
|
| 1456 |
+
"logits/rejected": -3.046875,
|
| 1457 |
+
"logps/chosen": -298.0,
|
| 1458 |
+
"logps/rejected": -616.0,
|
| 1459 |
+
"loss": 0.0,
|
| 1460 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1461 |
+
"memory_allocated (GB)": 51.7,
|
| 1462 |
+
"rewards/accuracies": 1.0,
|
| 1463 |
+
"rewards/chosen": -3.03125,
|
| 1464 |
+
"rewards/margins": 25.5,
|
| 1465 |
+
"rewards/rejected": -28.5,
|
| 1466 |
+
"step": 810,
|
| 1467 |
+
"total_memory_available (GB)": 94.62
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 4.21,
|
| 1471 |
+
"grad_norm": 0.0009728537406772375,
|
| 1472 |
+
"learning_rate": 4.7745751406263163e-05,
|
| 1473 |
+
"logits/chosen": -3.078125,
|
| 1474 |
+
"logits/rejected": -3.03125,
|
| 1475 |
+
"logps/chosen": -300.0,
|
| 1476 |
+
"logps/rejected": -620.0,
|
| 1477 |
+
"loss": 0.0001,
|
| 1478 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1479 |
+
"memory_allocated (GB)": 51.7,
|
| 1480 |
+
"rewards/accuracies": 1.0,
|
| 1481 |
+
"rewards/chosen": -3.125,
|
| 1482 |
+
"rewards/margins": 25.875,
|
| 1483 |
+
"rewards/rejected": -29.0,
|
| 1484 |
+
"step": 820,
|
| 1485 |
+
"total_memory_available (GB)": 94.62
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 4.26,
|
| 1489 |
+
"grad_norm": 0.0001809960085665807,
|
| 1490 |
+
"learning_rate": 4.274060686123959e-05,
|
| 1491 |
+
"logits/chosen": -3.109375,
|
| 1492 |
+
"logits/rejected": -3.03125,
|
| 1493 |
+
"logps/chosen": -300.0,
|
| 1494 |
+
"logps/rejected": -616.0,
|
| 1495 |
+
"loss": 0.0,
|
| 1496 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1497 |
+
"memory_allocated (GB)": 51.7,
|
| 1498 |
+
"rewards/accuracies": 1.0,
|
| 1499 |
+
"rewards/chosen": -2.84375,
|
| 1500 |
+
"rewards/margins": 25.375,
|
| 1501 |
+
"rewards/rejected": -28.25,
|
| 1502 |
+
"step": 830,
|
| 1503 |
+
"total_memory_available (GB)": 94.62
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 4.32,
|
| 1507 |
+
"grad_norm": 0.0019057235913351178,
|
| 1508 |
+
"learning_rate": 3.798797596089351e-05,
|
| 1509 |
+
"logits/chosen": -3.09375,
|
| 1510 |
+
"logits/rejected": -3.046875,
|
| 1511 |
+
"logps/chosen": -310.0,
|
| 1512 |
+
"logps/rejected": -620.0,
|
| 1513 |
+
"loss": 0.0,
|
| 1514 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1515 |
+
"memory_allocated (GB)": 51.7,
|
| 1516 |
+
"rewards/accuracies": 1.0,
|
| 1517 |
+
"rewards/chosen": -3.09375,
|
| 1518 |
+
"rewards/margins": 25.5,
|
| 1519 |
+
"rewards/rejected": -28.5,
|
| 1520 |
+
"step": 840,
|
| 1521 |
+
"total_memory_available (GB)": 94.62
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 4.37,
|
| 1525 |
+
"grad_norm": 0.007396820466965437,
|
| 1526 |
+
"learning_rate": 3.3493649053890325e-05,
|
| 1527 |
+
"logits/chosen": -3.078125,
|
| 1528 |
+
"logits/rejected": -3.015625,
|
| 1529 |
+
"logps/chosen": -310.0,
|
| 1530 |
+
"logps/rejected": -600.0,
|
| 1531 |
+
"loss": 0.0,
|
| 1532 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1533 |
+
"memory_allocated (GB)": 51.7,
|
| 1534 |
+
"rewards/accuracies": 1.0,
|
| 1535 |
+
"rewards/chosen": -3.203125,
|
| 1536 |
+
"rewards/margins": 24.625,
|
| 1537 |
+
"rewards/rejected": -27.75,
|
| 1538 |
+
"step": 850,
|
| 1539 |
+
"total_memory_available (GB)": 94.62
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 4.42,
|
| 1543 |
+
"grad_norm": 0.007374617271125317,
|
| 1544 |
+
"learning_rate": 2.9263101785268254e-05,
|
| 1545 |
+
"logits/chosen": -3.078125,
|
| 1546 |
+
"logits/rejected": -3.03125,
|
| 1547 |
+
"logps/chosen": -300.0,
|
| 1548 |
+
"logps/rejected": -616.0,
|
| 1549 |
+
"loss": 0.0001,
|
| 1550 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1551 |
+
"memory_allocated (GB)": 51.7,
|
| 1552 |
+
"rewards/accuracies": 1.0,
|
| 1553 |
+
"rewards/chosen": -3.03125,
|
| 1554 |
+
"rewards/margins": 25.5,
|
| 1555 |
+
"rewards/rejected": -28.5,
|
| 1556 |
+
"step": 860,
|
| 1557 |
+
"total_memory_available (GB)": 94.62
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 4.47,
|
| 1561 |
+
"grad_norm": 0.004346287343651056,
|
| 1562 |
+
"learning_rate": 2.5301488425208295e-05,
|
| 1563 |
+
"logits/chosen": -3.09375,
|
| 1564 |
+
"logits/rejected": -3.03125,
|
| 1565 |
+
"logps/chosen": -310.0,
|
| 1566 |
+
"logps/rejected": -624.0,
|
| 1567 |
+
"loss": 0.0,
|
| 1568 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1569 |
+
"memory_allocated (GB)": 51.7,
|
| 1570 |
+
"rewards/accuracies": 1.0,
|
| 1571 |
+
"rewards/chosen": -3.359375,
|
| 1572 |
+
"rewards/margins": 25.5,
|
| 1573 |
+
"rewards/rejected": -28.75,
|
| 1574 |
+
"step": 870,
|
| 1575 |
+
"total_memory_available (GB)": 94.62
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 4.52,
|
| 1579 |
+
"grad_norm": 0.0016928453696891665,
|
| 1580 |
+
"learning_rate": 2.1613635589349755e-05,
|
| 1581 |
+
"logits/chosen": -3.078125,
|
| 1582 |
+
"logits/rejected": -3.015625,
|
| 1583 |
+
"logps/chosen": -318.0,
|
| 1584 |
+
"logps/rejected": -620.0,
|
| 1585 |
+
"loss": 0.0,
|
| 1586 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1587 |
+
"memory_allocated (GB)": 51.7,
|
| 1588 |
+
"rewards/accuracies": 1.0,
|
| 1589 |
+
"rewards/chosen": -3.15625,
|
| 1590 |
+
"rewards/margins": 25.625,
|
| 1591 |
+
"rewards/rejected": -28.75,
|
| 1592 |
+
"step": 880,
|
| 1593 |
+
"total_memory_available (GB)": 94.62
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 4.57,
|
| 1597 |
+
"grad_norm": 0.0004315236583352089,
|
| 1598 |
+
"learning_rate": 1.8204036358303172e-05,
|
| 1599 |
+
"logits/chosen": -3.09375,
|
| 1600 |
+
"logits/rejected": -3.046875,
|
| 1601 |
+
"logps/chosen": -302.0,
|
| 1602 |
+
"logps/rejected": -608.0,
|
| 1603 |
+
"loss": 0.0,
|
| 1604 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1605 |
+
"memory_allocated (GB)": 51.7,
|
| 1606 |
+
"rewards/accuracies": 1.0,
|
| 1607 |
+
"rewards/chosen": -3.171875,
|
| 1608 |
+
"rewards/margins": 25.125,
|
| 1609 |
+
"rewards/rejected": -28.375,
|
| 1610 |
+
"step": 890,
|
| 1611 |
+
"total_memory_available (GB)": 94.62
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 4.62,
|
| 1615 |
+
"grad_norm": 0.0036792384926229715,
|
| 1616 |
+
"learning_rate": 1.5076844803522921e-05,
|
| 1617 |
+
"logits/chosen": -3.09375,
|
| 1618 |
+
"logits/rejected": -3.046875,
|
| 1619 |
+
"logps/chosen": -322.0,
|
| 1620 |
+
"logps/rejected": -624.0,
|
| 1621 |
+
"loss": 0.0,
|
| 1622 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1623 |
+
"memory_allocated (GB)": 51.7,
|
| 1624 |
+
"rewards/accuracies": 1.0,
|
| 1625 |
+
"rewards/chosen": -3.28125,
|
| 1626 |
+
"rewards/margins": 25.375,
|
| 1627 |
+
"rewards/rejected": -28.75,
|
| 1628 |
+
"step": 900,
|
| 1629 |
+
"total_memory_available (GB)": 94.62
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 4.68,
|
| 1633 |
+
"grad_norm": 0.0032357031013816595,
|
| 1634 |
+
"learning_rate": 1.2235870926211617e-05,
|
| 1635 |
+
"logits/chosen": -3.09375,
|
| 1636 |
+
"logits/rejected": -3.03125,
|
| 1637 |
+
"logps/chosen": -308.0,
|
| 1638 |
+
"logps/rejected": -608.0,
|
| 1639 |
+
"loss": 0.0001,
|
| 1640 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1641 |
+
"memory_allocated (GB)": 51.7,
|
| 1642 |
+
"rewards/accuracies": 1.0,
|
| 1643 |
+
"rewards/chosen": -3.0,
|
| 1644 |
+
"rewards/margins": 25.125,
|
| 1645 |
+
"rewards/rejected": -28.125,
|
| 1646 |
+
"step": 910,
|
| 1647 |
+
"total_memory_available (GB)": 94.62
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 4.73,
|
| 1651 |
+
"grad_norm": 0.00263414834626019,
|
| 1652 |
+
"learning_rate": 9.684576015420277e-06,
|
| 1653 |
+
"logits/chosen": -3.09375,
|
| 1654 |
+
"logits/rejected": -3.03125,
|
| 1655 |
+
"logps/chosen": -308.0,
|
| 1656 |
+
"logps/rejected": -608.0,
|
| 1657 |
+
"loss": 0.0001,
|
| 1658 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1659 |
+
"memory_allocated (GB)": 51.7,
|
| 1660 |
+
"rewards/accuracies": 1.0,
|
| 1661 |
+
"rewards/chosen": -3.265625,
|
| 1662 |
+
"rewards/margins": 25.375,
|
| 1663 |
+
"rewards/rejected": -28.625,
|
| 1664 |
+
"step": 920,
|
| 1665 |
+
"total_memory_available (GB)": 94.62
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 4.78,
|
| 1669 |
+
"grad_norm": 0.0006815909291617572,
|
| 1670 |
+
"learning_rate": 7.426068431000882e-06,
|
| 1671 |
+
"logits/chosen": -3.09375,
|
| 1672 |
+
"logits/rejected": -3.03125,
|
| 1673 |
+
"logps/chosen": -300.0,
|
| 1674 |
+
"logps/rejected": -612.0,
|
| 1675 |
+
"loss": 0.0001,
|
| 1676 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1677 |
+
"memory_allocated (GB)": 51.7,
|
| 1678 |
+
"rewards/accuracies": 1.0,
|
| 1679 |
+
"rewards/chosen": -3.171875,
|
| 1680 |
+
"rewards/margins": 25.0,
|
| 1681 |
+
"rewards/rejected": -28.25,
|
| 1682 |
+
"step": 930,
|
| 1683 |
+
"total_memory_available (GB)": 94.62
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 4.83,
|
| 1687 |
+
"grad_norm": 0.007029928732663393,
|
| 1688 |
+
"learning_rate": 5.463099816548578e-06,
|
| 1689 |
+
"logits/chosen": -3.109375,
|
| 1690 |
+
"logits/rejected": -3.046875,
|
| 1691 |
+
"logps/chosen": -314.0,
|
| 1692 |
+
"logps/rejected": -624.0,
|
| 1693 |
+
"loss": 0.0001,
|
| 1694 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1695 |
+
"memory_allocated (GB)": 51.7,
|
| 1696 |
+
"rewards/accuracies": 1.0,
|
| 1697 |
+
"rewards/chosen": -3.28125,
|
| 1698 |
+
"rewards/margins": 26.0,
|
| 1699 |
+
"rewards/rejected": -29.25,
|
| 1700 |
+
"step": 940,
|
| 1701 |
+
"total_memory_available (GB)": 94.62
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 4.88,
|
| 1705 |
+
"grad_norm": 0.0016696957172825933,
|
| 1706 |
+
"learning_rate": 3.798061746947995e-06,
|
| 1707 |
+
"logits/chosen": -3.078125,
|
| 1708 |
+
"logits/rejected": -3.03125,
|
| 1709 |
+
"logps/chosen": -292.0,
|
| 1710 |
+
"logps/rejected": -616.0,
|
| 1711 |
+
"loss": 0.0,
|
| 1712 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1713 |
+
"memory_allocated (GB)": 51.7,
|
| 1714 |
+
"rewards/accuracies": 1.0,
|
| 1715 |
+
"rewards/chosen": -3.21875,
|
| 1716 |
+
"rewards/margins": 25.25,
|
| 1717 |
+
"rewards/rejected": -28.375,
|
| 1718 |
+
"step": 950,
|
| 1719 |
+
"total_memory_available (GB)": 94.62
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 4.93,
|
| 1723 |
+
"grad_norm": 0.004225951619446278,
|
| 1724 |
+
"learning_rate": 2.4329828146074094e-06,
|
| 1725 |
+
"logits/chosen": -3.09375,
|
| 1726 |
+
"logits/rejected": -3.046875,
|
| 1727 |
+
"logps/chosen": -308.0,
|
| 1728 |
+
"logps/rejected": -624.0,
|
| 1729 |
+
"loss": 0.0,
|
| 1730 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1731 |
+
"memory_allocated (GB)": 51.7,
|
| 1732 |
+
"rewards/accuracies": 1.0,
|
| 1733 |
+
"rewards/chosen": -3.203125,
|
| 1734 |
+
"rewards/margins": 26.0,
|
| 1735 |
+
"rewards/rejected": -29.125,
|
| 1736 |
+
"step": 960,
|
| 1737 |
+
"total_memory_available (GB)": 94.62
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 4.98,
|
| 1741 |
+
"grad_norm": 0.00014331191778182983,
|
| 1742 |
+
"learning_rate": 1.3695261579316775e-06,
|
| 1743 |
+
"logits/chosen": -3.09375,
|
| 1744 |
+
"logits/rejected": -3.015625,
|
| 1745 |
+
"logps/chosen": -312.0,
|
| 1746 |
+
"logps/rejected": -604.0,
|
| 1747 |
+
"loss": 0.0001,
|
| 1748 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1749 |
+
"memory_allocated (GB)": 51.7,
|
| 1750 |
+
"rewards/accuracies": 1.0,
|
| 1751 |
+
"rewards/chosen": -3.15625,
|
| 1752 |
+
"rewards/margins": 25.5,
|
| 1753 |
+
"rewards/rejected": -28.75,
|
| 1754 |
+
"step": 970,
|
| 1755 |
+
"total_memory_available (GB)": 94.62
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 5.04,
|
| 1759 |
+
"grad_norm": 0.0036776724737137556,
|
| 1760 |
+
"learning_rate": 6.089874350439506e-07,
|
| 1761 |
+
"logits/chosen": -3.09375,
|
| 1762 |
+
"logits/rejected": -3.046875,
|
| 1763 |
+
"logps/chosen": -330.0,
|
| 1764 |
+
"logps/rejected": -616.0,
|
| 1765 |
+
"loss": 0.0001,
|
| 1766 |
+
"max_memory_allocated (GB)": 94.39,
|
| 1767 |
+
"memory_allocated (GB)": 51.7,
|
| 1768 |
+
"rewards/accuracies": 1.0,
|
| 1769 |
+
"rewards/chosen": -3.3125,
|
| 1770 |
+
"rewards/margins": 25.25,
|
| 1771 |
+
"rewards/rejected": -28.5,
|
| 1772 |
+
"step": 980,
|
| 1773 |
+
"total_memory_available (GB)": 94.62
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 5.09,
|
| 1777 |
+
"grad_norm": 1.2631918252736796e-05,
|
| 1778 |
+
"learning_rate": 1.5229324522605948e-07,
|
| 1779 |
+
"logits/chosen": -3.09375,
|
| 1780 |
+
"logits/rejected": -3.046875,
|
| 1781 |
+
"logps/chosen": -300.0,
|
| 1782 |
+
"logps/rejected": -608.0,
|
| 1783 |
+
"loss": 0.0,
|
| 1784 |
+
"max_memory_allocated (GB)": 94.42,
|
| 1785 |
+
"memory_allocated (GB)": 51.7,
|
| 1786 |
+
"rewards/accuracies": 1.0,
|
| 1787 |
+
"rewards/chosen": -3.328125,
|
| 1788 |
+
"rewards/margins": 25.0,
|
| 1789 |
+
"rewards/rejected": -28.375,
|
| 1790 |
+
"step": 990,
|
| 1791 |
+
"total_memory_available (GB)": 94.62
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 5.14,
|
| 1795 |
+
"grad_norm": 0.0053689004853367805,
|
| 1796 |
+
"learning_rate": 0.0,
|
| 1797 |
+
"logits/chosen": -3.09375,
|
| 1798 |
+
"logits/rejected": -3.0625,
|
| 1799 |
+
"logps/chosen": -304.0,
|
| 1800 |
+
"logps/rejected": -624.0,
|
| 1801 |
+
"loss": 0.0,
|
| 1802 |
+
"max_memory_allocated (GB)": 94.42,
|
| 1803 |
+
"memory_allocated (GB)": 51.7,
|
| 1804 |
+
"rewards/accuracies": 1.0,
|
| 1805 |
+
"rewards/chosen": -3.265625,
|
| 1806 |
+
"rewards/margins": 25.25,
|
| 1807 |
+
"rewards/rejected": -28.625,
|
| 1808 |
+
"step": 1000,
|
| 1809 |
+
"total_memory_available (GB)": 94.62
|
| 1810 |
+
}
|
| 1811 |
+
],
|
| 1812 |
+
"logging_steps": 10,
|
| 1813 |
+
"max_steps": 1000,
|
| 1814 |
+
"num_input_tokens_seen": 0,
|
| 1815 |
+
"num_train_epochs": 6,
|
| 1816 |
+
"save_steps": 10,
|
| 1817 |
+
"total_flos": 8.39920851615744e+18,
|
| 1818 |
+
"train_batch_size": 8,
|
| 1819 |
+
"trial_name": null,
|
| 1820 |
+
"trial_params": null
|
| 1821 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ecb3ba82f8c3d2be70d87ab7b4cc65a3123c2080711af38653cafe8daf92937
|
| 3 |
+
size 4728
|
checkpoint-110/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.4.0
|
checkpoint-110/adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "./neuralchat_v3_finetuned_model_lora",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"k_proj",
|
| 18 |
+
"q_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
checkpoint-110/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1518e0ecda9f14d0773d07778c6df1e690f867b0c584e5ffbdeafc02678ef1da
|
| 3 |
+
size 37774528
|
checkpoint-110/gaudi_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"autocast_bf16_ops": null,
|
| 3 |
+
"autocast_fp32_ops": null,
|
| 4 |
+
"optimum_version": "1.19.2",
|
| 5 |
+
"transformers_version": "4.38.1",
|
| 6 |
+
"use_dynamic_shapes": false,
|
| 7 |
+
"use_fused_adam": true,
|
| 8 |
+
"use_fused_clip_norm": true,
|
| 9 |
+
"use_torch_autocast": false
|
| 10 |
+
}
|
checkpoint-110/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1781d7c3c97a725739d5d6e736a718a737ac8a0ec48427a2801ec98c998b4ca6
|
| 3 |
+
size 75593978
|
checkpoint-110/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:880d021a68f50eef9ea2c2d3b13e7b2237662239383d04734837f5bf6eb202da
|
| 3 |
+
size 18020
|
checkpoint-110/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b71962485900b43ccfba6f9b2f08c7975fe312975bb825b2c328bc2adbe8f3a
|
| 3 |
+
size 1000
|
checkpoint-110/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|