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-rabid_bristly_elephant
 tags:
@@ -14,7 +14,7 @@ licence: license
 
 # Model Card for Qwen2.5-0.5B-Instruct-Gensyn-Swarm-rabid_bristly_elephant
 
-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": 1.1510408342019218e-06,
-    "train_runtime": 8889.6683,
-    "train_samples": 17,
-    "train_samples_per_second": 0.018,
+    "train_loss": 5.6818393204594034e-06,
+    "train_runtime": 3692.0926,
+    "train_samples": 4,
+    "train_samples_per_second": 0.005,
     "train_steps_per_second": 0.001
 }

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 sha256:8ccf081049c4862547cff9463bad956ff6aca0b527647cdd03541a14f54f1c71
+oid sha256:47f0b98019fab60d917842d7630232b45b0707a6f9e8bffad2c0a810d275c07a
 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": 1.1510408342019218e-06,
-    "train_runtime": 8889.6683,
-    "train_samples": 17,
-    "train_samples_per_second": 0.018,
+    "train_loss": 5.6818393204594034e-06,
+    "train_runtime": 3692.0926,
+    "train_samples": 4,
+    "train_samples_per_second": 0.005,
     "train_steps_per_second": 0.001
 }

trainer_state.json

@@ -2,123 +2,78 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.4705882352941178,
+  "epoch": 2.5,
   "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": 244.4375,
-      "epoch": 0.9411764705882353,
-      "grad_norm": 0.9346978068351746,
+      "completion_length": 140.125,
+      "epoch": 1.0,
+      "grad_norm": 25.24424934387207,
       "kl": 0.0,
       "learning_rate": 5e-07,
-      "loss": -0.0,
-      "reward": 6.4995622634887695,
-      "reward_std": 0.3433886617422104,
-      "rewards/concensus_correctness_reward_func": 1.9240000247955322,
-      "rewards/consensus_reward_func": 2.0,
-      "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5,
-      "rewards/question_recreation_reward_func": 1.0,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 1.0755624920129776,
-      "step": 2
-    },
-    {
-      "completion_length": 243.0,
-      "epoch": 1.4705882352941178,
-      "grad_norm": 1.7506262063980103,
-      "kl": 0.0007122650648751813,
-      "learning_rate": 4.415111107797445e-07,
       "loss": 0.0,
-      "reward": 6.252771430545383,
-      "reward_std": 0.596649514304267,
-      "rewards/concensus_correctness_reward_func": 1.9240000247955322,
-      "rewards/consensus_reward_func": 2.0,
+      "reward": 0.052890727296471596,
+      "reward_std": 0.017691312765236944,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.3333333333333333,
-      "rewards/question_recreation_reward_func": 0.9508272144529555,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.052890727296471596,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 1.0446111030048795,
-      "step": 4
+      "rewards/xmlcount_reward_func": 0.0,
+      "step": 2
     },
     {
-      "completion_length": 280.27777777777777,
+      "completion_length": 86.625,
       "epoch": 2.0,
-      "grad_norm": 0.6869510412216187,
-      "kl": 0.0011853337701823977,
-      "learning_rate": 2.934120444167326e-07,
-      "loss": 0.0,
-      "reward": 6.119110955132379,
-      "reward_std": 0.15713479783799914,
-      "rewards/concensus_correctness_reward_func": 1.9240000247955322,
-      "rewards/consensus_reward_func": 2.0,
-      "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1111111111111111,
-      "rewards/question_recreation_reward_func": 1.0,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 1.0839999914169312,
-      "step": 6
-    },
-    {
-      "completion_length": 283.21875,
-      "epoch": 2.9411764705882355,
-      "grad_norm": 1.2507210969924927,
-      "kl": 0.003197207028279081,
-      "learning_rate": 1.2500000000000005e-07,
+      "grad_norm": 9.793798446655273,
+      "kl": 0.011194768259883858,
+      "learning_rate": 2.5e-07,
       "loss": 0.0,
-      "reward": 6.257999777793884,
-      "reward_std": 0.35355329513549805,
-      "rewards/concensus_correctness_reward_func": 1.9240000247955322,
-      "rewards/consensus_reward_func": 2.0,
+      "reward": 0.10395016614347696,
+      "reward_std": 0.012995281023904681,
+      "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": 1.0,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.10395016614347696,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 1.0839999914169312,
-      "step": 8
+      "rewards/xmlcount_reward_func": 0.0,
+      "step": 4
     },
     {
-      "completion_length": 280.94444444444446,
-      "epoch": 3.4705882352941178,
-      "grad_norm": 1.148160696029663,
-      "kl": 0.002885583956311974,
-      "learning_rate": 1.507684480352292e-08,
-      "loss": 0.0,
-      "reward": 6.230222013261583,
-      "reward_std": 0.31426959567599827,
-      "rewards/concensus_correctness_reward_func": 1.9240000247955322,
-      "rewards/consensus_reward_func": 2.0,
+      "completion_length": 104.0,
+      "epoch": 2.5,
+      "kl": 0.005378011410357431,
+      "reward": 0.01932604657486081,
+      "reward_std": 0.00301999697512656,
+      "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": 1.0,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.01932604657486081,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 1.0839999914169312,
-      "step": 10
-    },
-    {
-      "epoch": 3.4705882352941178,
-      "step": 10,
+      "rewards/xmlcount_reward_func": 0.0,
+      "step": 5,
       "total_flos": 0.0,
-      "train_loss": 1.1510408342019218e-06,
-      "train_runtime": 8889.6683,
-      "train_samples_per_second": 0.018,
+      "train_loss": 5.6818393204594034e-06,
+      "train_runtime": 3692.0926,
+      "train_samples_per_second": 0.005,
       "train_steps_per_second": 0.001
     }
   ],
   "logging_steps": 2,
-  "max_steps": 10,
+  "max_steps": 5,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
-  "save_steps": 25,
+  "num_train_epochs": 3,
+  "save_steps": 10,
   "stateful_callbacks": {
     "TrainerControl": {
       "args": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa8d017a40052868c6883eba594a187b92e66b3749716bc140436bffe854bead
+oid sha256:8da921a3c6b3ad2119ef1e9db6976fc1c39437f577354c320f482b2ce4da1e4f
 size 5944