End of training

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: Gensyn/Qwen2.5-0.5B-Instruct
+base_model: unsloth/Qwen2.5-0.5B-Instruct
 library_name: transformers
 model_name: Qwen2.5-0.5B-Instruct-Gensyn-Swarm-lethal_tawny_deer
 tags:
@@ -14,7 +14,7 @@ licence: license
 
 # Model Card for Qwen2.5-0.5B-Instruct-Gensyn-Swarm-lethal_tawny_deer
 
-This model is a fine-tuned version of [Gensyn/Qwen2.5-0.5B-Instruct](https://huggingface.co/Gensyn/Qwen2.5-0.5B-Instruct).
+This model is a fine-tuned version of [unsloth/Qwen2.5-0.5B-Instruct](https://huggingface.co/unsloth/Qwen2.5-0.5B-Instruct).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 4.9965383652761375e-06,
-    "train_runtime": 5282.5247,
-    "train_samples": 17,
-    "train_samples_per_second": 0.03,
-    "train_steps_per_second": 0.002
+    "train_loss": 7.331371307373047e-07,
+    "train_runtime": 1043.6808,
+    "train_samples": 3,
+    "train_samples_per_second": 0.019,
+    "train_steps_per_second": 0.005
 }

config.json

@@ -3,7 +3,6 @@
     "Qwen2ForCausalLM"
   ],
   "attention_dropout": 0.0,
-  "bos_token_id": 151643,
   "eos_token_id": 151645,
   "hidden_act": "silu",
   "hidden_size": 896,
@@ -15,13 +14,15 @@
   "num_attention_heads": 14,
   "num_hidden_layers": 24,
   "num_key_value_heads": 2,
+  "pad_token_id": 151654,
   "rms_norm_eps": 1e-06,
   "rope_scaling": null,
   "rope_theta": 1000000.0,
-  "sliding_window": 32768,
+  "sliding_window": null,
   "tie_word_embeddings": true,
   "torch_dtype": "float32",
   "transformers_version": "4.51.2",
+  "unsloth_fixed": true,
   "use_cache": true,
   "use_sliding_window": false,
   "vocab_size": 151936

generation_config.json

@@ -5,7 +5,8 @@
     151645,
     151643
   ],
-  "pad_token_id": 151643,
+  "max_length": 32768,
+  "pad_token_id": 151654,
   "repetition_penalty": 1.1,
   "temperature": 0.7,
   "top_k": 20,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6fe9160b49ab9da979f8e5a9e1b761aa0608f531a05651ad604686e8252caf69
+oid sha256:e573dc601c408ba8b5ae9d6c63e6cb6686e6332c73231c4810ba5c2bf425bfac
 size 1976163472

special_tokens_map.json

@@ -22,7 +22,7 @@
     "single_word": false
   },
   "pad_token": {
-    "content": "<|endoftext|>",
+    "content": "<|vision_pad|>",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5eee858c5123a4279c3e1f7b81247343f356ac767940b2692a928ad929543214
-size 11422063
+oid sha256:64e71213db910f5cafa86d35091f37393dcc344b1bbc34091d1b3eed4cca01d5
+size 11422064

tokenizer_config.json

@@ -200,8 +200,9 @@
   "eos_token": "<|im_end|>",
   "errors": "replace",
   "extra_special_tokens": {},
-  "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
+  "model_max_length": 32768,
+  "pad_token": "<|vision_pad|>",
+  "padding_side": "left",
   "split_special_tokens": false,
   "tokenizer_class": "Qwen2Tokenizer",
   "unk_token": null

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 4.9965383652761375e-06,
-    "train_runtime": 5282.5247,
-    "train_samples": 17,
-    "train_samples_per_second": 0.03,
-    "train_steps_per_second": 0.002
+    "train_loss": 7.331371307373047e-07,
+    "train_runtime": 1043.6808,
+    "train_samples": 3,
+    "train_samples_per_second": 0.019,
+    "train_steps_per_second": 0.005
 }

trainer_state.json

@@ -2,123 +2,78 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.4705882352941178,
+  "epoch": 2.6666666666666665,
   "eval_steps": 500,
-  "global_step": 10,
+  "global_step": 5,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 105.875,
-      "epoch": 0.9411764705882353,
-      "grad_norm": 0.0,
+      "completion_length": 127.16666666666667,
+      "epoch": 1.0,
+      "grad_norm": 10.25474739074707,
       "kl": 0.0,
       "learning_rate": 5e-07,
-      "loss": 0.0,
-      "reward": 7.015919391065836,
-      "reward_std": 0.010885396040976048,
-      "rewards/concensus_correctness_reward_func": 2.176249995827675,
-      "rewards/consensus_reward_func": 1.875,
+      "loss": -0.0,
+      "reward": 0.16262217486898103,
+      "reward_std": 0.08637613849714398,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.9412319990806282,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.16262217486898103,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.46875,
-      "rewards/xmlcount_reward_func": 1.1796875,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.0,
       "step": 2
     },
     {
-      "completion_length": 104.61111111111111,
-      "epoch": 1.4705882352941178,
-      "grad_norm": 5.037790775299072,
-      "kl": 0.0018976248730445554,
-      "learning_rate": 4.415111107797445e-07,
-      "loss": 0.0,
-      "reward": 6.546576562854979,
-      "reward_std": 0.0986187085768001,
-      "rewards/concensus_correctness_reward_func": 2.162222226460775,
-      "rewards/consensus_reward_func": 1.5555555555555556,
-      "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.6666666666666666,
-      "rewards/question_recreation_reward_func": 0.7940766894155078,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.3611111111111111,
-      "rewards/xmlcount_reward_func": 1.0069444444444444,
-      "step": 4
-    },
-    {
-      "completion_length": 121.55555555555556,
+      "completion_length": 187.5,
       "epoch": 2.0,
-      "grad_norm": 1.8580118417739868,
-      "kl": 0.0006738082559220655,
-      "learning_rate": 2.934120444167326e-07,
+      "grad_norm": 8.673498153686523,
+      "kl": 0.0010691387384819488,
+      "learning_rate": 2.5e-07,
       "loss": 0.0,
-      "reward": 5.957985914415783,
-      "reward_std": 0.022622719407081604,
-      "rewards/concensus_correctness_reward_func": 1.7066666550106473,
-      "rewards/consensus_reward_func": 1.7777777777777777,
+      "reward": 0.10594786455233891,
+      "reward_std": 0.033774735095600285,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.9040971166557736,
+      "rewards/question_recreation_reward_func": 0.10594786455233891,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.4444444444444444,
-      "rewards/xmlcount_reward_func": 1.125,
-      "step": 6
-    },
-    {
-      "completion_length": 95.625,
-      "epoch": 2.9411764705882355,
-      "grad_norm": 3.880513906478882,
-      "kl": 0.013746480780241654,
-      "learning_rate": 1.2500000000000005e-07,
-      "loss": 0.0,
-      "reward": 6.583770731464028,
-      "reward_std": 0.0842309154686518,
-      "rewards/concensus_correctness_reward_func": 2.056499995291233,
-      "rewards/consensus_reward_func": 1.75,
-      "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.8952083382755518,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.40625,
-      "rewards/xmlcount_reward_func": 1.1008124947547913,
-      "step": 8
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.0,
+      "step": 4
     },
     {
-      "completion_length": 106.05555555555556,
-      "epoch": 3.4705882352941178,
-      "grad_norm": 3.5325653553009033,
-      "kl": 0.015616837632579619,
-      "learning_rate": 1.507684480352292e-08,
-      "loss": 0.0,
-      "reward": 5.260379309455554,
-      "reward_std": 0.027002629927463002,
-      "rewards/concensus_correctness_reward_func": 1.4937777784135606,
-      "rewards/consensus_reward_func": 1.5555555555555556,
+      "completion_length": 126.25,
+      "epoch": 2.6666666666666665,
+      "kl": 0.0008855987689457834,
+      "reward": 0.1095019057393074,
+      "reward_std": 0.005491654243087396,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.808268278837204,
+      "rewards/question_recreation_reward_func": 0.1095019057393074,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.3888888888888889,
-      "rewards/xmlcount_reward_func": 1.0138888888888888,
-      "step": 10
-    },
-    {
-      "epoch": 3.4705882352941178,
-      "step": 10,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.0,
+      "step": 5,
       "total_flos": 0.0,
-      "train_loss": 4.9965383652761375e-06,
-      "train_runtime": 5282.5247,
-      "train_samples_per_second": 0.03,
-      "train_steps_per_second": 0.002
+      "train_loss": 7.331371307373047e-07,
+      "train_runtime": 1043.6808,
+      "train_samples_per_second": 0.019,
+      "train_steps_per_second": 0.005
     }
   ],
   "logging_steps": 2,
-  "max_steps": 10,
+  "max_steps": 5,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 5,
-  "save_steps": 25,
+  "save_steps": 10,
   "stateful_callbacks": {
     "TrainerControl": {
       "args": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:69815f1dc525e3403e905adf7230c3cee33a2c72d178fea037f37dbff247a476
+oid sha256:a6e8bda82286f4ad9312e40a24819e6419d65b09f5d5840cb46a13ab0b6e2aa3
 size 5944