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-gilded_shrewd_gecko
 tags:
@@ -14,7 +14,7 @@ licence: license
 
 # Model Card for Qwen2.5-0.5B-Instruct-Gensyn-Swarm-gilded_shrewd_gecko
 
-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": 2.334083110326901e-05,
-    "train_runtime": 17913.3066,
-    "train_samples": 17,
-    "train_samples_per_second": 0.009,
-    "train_steps_per_second": 0.001
+    "train_loss": 5.349516868591308e-07,
+    "train_runtime": 33763.7805,
+    "train_samples": 5,
+    "train_samples_per_second": 0.001,
+    "train_steps_per_second": 0.0
 }

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 sha256:15d8162ff2f5da1973710d93f2b47430e7ce8999fe19ac9f1612de6271afd4b5
+oid sha256:d994cea4e26e67b289f513b952d75f515fa334b90e8ab27ed9834f0453de6a69
 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": 2.334083110326901e-05,
-    "train_runtime": 17913.3066,
-    "train_samples": 17,
-    "train_samples_per_second": 0.009,
-    "train_steps_per_second": 0.001
+    "train_loss": 5.349516868591308e-07,
+    "train_runtime": 33763.7805,
+    "train_samples": 5,
+    "train_samples_per_second": 0.001,
+    "train_steps_per_second": 0.0
 }

trainer_state.json

@@ -2,120 +2,75 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.4705882352941178,
+  "epoch": 2.8,
   "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": 131.78125,
-      "epoch": 0.9411764705882353,
-      "grad_norm": 10.730029106140137,
+      "completion_length": 191.7,
+      "epoch": 1.0,
+      "grad_norm": 3.4990100860595703,
       "kl": 0.0,
       "learning_rate": 5e-07,
-      "loss": -0.0,
-      "reward": 6.342288956046104,
-      "reward_std": 0.6563283740542829,
-      "rewards/concensus_correctness_reward_func": 2.102062474936247,
-      "rewards/consensus_reward_func": 1.9375,
-      "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.25,
-      "rewards/question_recreation_reward_func": 0.9382578537333757,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.15625,
-      "rewards/xmlcount_reward_func": 0.9582187570631504,
-      "step": 2
-    },
-    {
-      "completion_length": 93.83333333333333,
-      "epoch": 1.4705882352941178,
-      "grad_norm": 7.171228408813477,
-      "kl": 0.017023057715656858,
-      "learning_rate": 4.415111107797445e-07,
       "loss": 0.0,
-      "reward": 6.55574173397488,
-      "reward_std": 0.43832636404679054,
-      "rewards/concensus_correctness_reward_func": 2.472111072805193,
-      "rewards/consensus_reward_func": 1.8888888888888888,
+      "reward": 0.13842474818229675,
+      "reward_std": 0.09636008515954017,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4444444444444444,
-      "rewards/question_recreation_reward_func": 0.7796862405310903,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.11342475041747094,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.1111111111111111,
-      "rewards/xmlcount_reward_func": 0.8594999843173556,
-      "step": 4
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.025,
+      "step": 2
     },
     {
-      "completion_length": 122.61111111111111,
+      "completion_length": 173.0,
       "epoch": 2.0,
-      "grad_norm": 3.105441237494233e-08,
-      "kl": 0.02888714972262581,
-      "learning_rate": 2.934120444167326e-07,
-      "loss": 0.0,
-      "reward": 6.286137633853489,
-      "reward_std": 0.16779031413089898,
-      "rewards/concensus_correctness_reward_func": 2.134666668044196,
-      "rewards/consensus_reward_func": 2.0,
-      "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1111111111111111,
-      "rewards/question_recreation_reward_func": 0.8890821365122166,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.1111111111111111,
-      "rewards/xmlcount_reward_func": 1.0401666694217258,
-      "step": 6
-    },
-    {
-      "completion_length": 121.0625,
-      "epoch": 2.9411764705882355,
-      "grad_norm": 5.204808235168457,
-      "kl": 0.036964431335945847,
-      "learning_rate": 1.2500000000000005e-07,
+      "grad_norm": 2.3786094188690186,
+      "kl": 0.0006931740383151919,
+      "learning_rate": 2.5e-07,
       "loss": 0.0,
-      "reward": 6.445510894060135,
-      "reward_std": 0.12171002507966477,
-      "rewards/concensus_correctness_reward_func": 2.28712497651577,
-      "rewards/consensus_reward_func": 2.0,
+      "reward": 0.17476099207997323,
+      "reward_std": 0.059773775190114974,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.25,
-      "rewards/question_recreation_reward_func": 0.8760735102987383,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.1956609971821308,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.109375,
-      "rewards/xmlcount_reward_func": 0.9229374881833792,
-      "step": 8
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": -0.020899999886751175,
+      "step": 4
     },
     {
-      "completion_length": 135.33333333333334,
-      "epoch": 3.4705882352941178,
-      "grad_norm": 13.752861976623535,
-      "kl": 0.08299477315611309,
-      "learning_rate": 1.507684480352292e-08,
-      "loss": 0.0,
-      "reward": 6.4182407591078015,
-      "reward_std": 0.18086219827334085,
-      "rewards/concensus_correctness_reward_func": 2.3553333150015936,
-      "rewards/consensus_reward_func": 2.0,
+      "completion_length": 135.875,
+      "epoch": 2.8,
+      "kl": 0.0009647479673731141,
+      "reward": 0.09912843629717827,
+      "reward_std": 0.05026640766300261,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.2222222222222222,
-      "rewards/question_recreation_reward_func": 0.7799075724453561,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.0678784353658557,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.1111111111111111,
-      "rewards/xmlcount_reward_func": 0.9496666457917955,
-      "step": 10
-    },
-    {
-      "epoch": 3.4705882352941178,
-      "step": 10,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.03125,
+      "step": 5,
       "total_flos": 0.0,
-      "train_loss": 2.334083110326901e-05,
-      "train_runtime": 17913.3066,
-      "train_samples_per_second": 0.009,
-      "train_steps_per_second": 0.001
+      "train_loss": 5.349516868591308e-07,
+      "train_runtime": 33763.7805,
+      "train_samples_per_second": 0.001,
+      "train_steps_per_second": 0.0
     }
   ],
   "logging_steps": 2,
-  "max_steps": 10,
+  "max_steps": 5,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 5,
   "save_steps": 25,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:852aacd4d84d703d5046f7d86492b82ca4f7b4f2e2b19b3deda376ebc79b9afb
+oid sha256:d1abc1be2cf80e7a84ea2ff75490a1d1c8184d0360e57b98a7d0375d568a07be
 size 5944