Training in progress, step 23

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,69 @@
+---
+library_name: transformers
+license: other
+base_model: trl-lib/qwen1.5-0.5b-sft
+tags:
+- alignment-handbook
+- trl
+- simpo
+- generated_from_trainer
+model-index:
+- name: trl-lib/qwen1.5-0.5b-sft
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# trl-lib/qwen1.5-0.5b-sft
+
+This model is a fine-tuned version of [trl-lib/qwen1.5-0.5b-sft](https://huggingface.co/trl-lib/qwen1.5-0.5b-sft) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 6.4892
+- Rewards/chosen: -17.8732
+- Rewards/rejected: -18.7312
+- Rewards/accuracies: 0.5237
+- Rewards/margins: 0.8579
+- Logps/rejected: -1.8731
+- Logps/chosen: -1.7873
+- Logits/rejected: -0.3281
+- Logits/chosen: -0.3588
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 8e-08
+- train_batch_size: 1
+- eval_batch_size: 4
+- seed: 42
+- gradient_accumulation_steps: 16
+- total_train_batch_size: 16
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.1
+- training_steps: 20
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 4.48.3
+- Pytorch 2.5.1+cu124
+- Datasets 3.2.0
+- Tokenizers 0.21.0

added_tokens.json

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

all_results.json

@@ -0,0 +1,21 @@
+{
+    "epoch": 0.005887272119083459,
+    "eval_logits/chosen": -0.3587622344493866,
+    "eval_logits/rejected": -0.32808682322502136,
+    "eval_logps/chosen": -1.7873238325119019,
+    "eval_logps/rejected": -1.8731167316436768,
+    "eval_loss": 6.489209175109863,
+    "eval_rewards/accuracies": 0.5237388610839844,
+    "eval_rewards/chosen": -17.87323760986328,
+    "eval_rewards/margins": 0.8579314351081848,
+    "eval_rewards/rejected": -18.731168746948242,
+    "eval_runtime": 60.3757,
+    "eval_samples": 1345,
+    "eval_samples_per_second": 22.277,
+    "eval_steps_per_second": 5.582,
+    "total_flos": 0.0,
+    "train_loss": 4.74956651167436,
+    "train_runtime": 13.2482,
+    "train_samples_per_second": 24.154,
+    "train_steps_per_second": 1.51
+}

config.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "trl-lib/qwen1.5-0.5b-sft",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151644,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 16,
+  "pad_token_id": 151645,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.48.3",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151648
+}

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.005887272119083459,
+    "eval_logits/chosen": -0.3587622344493866,
+    "eval_logits/rejected": -0.32808682322502136,
+    "eval_logps/chosen": -1.7873238325119019,
+    "eval_logps/rejected": -1.8731167316436768,
+    "eval_loss": 6.489209175109863,
+    "eval_rewards/accuracies": 0.5237388610839844,
+    "eval_rewards/chosen": -17.87323760986328,
+    "eval_rewards/margins": 0.8579314351081848,
+    "eval_rewards/rejected": -18.731168746948242,
+    "eval_runtime": 60.3757,
+    "eval_samples": 1345,
+    "eval_samples_per_second": 22.277,
+    "eval_steps_per_second": 5.582
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "bos_token_id": 151644,
+  "eos_token_id": 151645,
+  "max_new_tokens": 2048,
+  "pad_token_id": 151645,
+  "transformers_version": "4.48.3"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa79e3a29fc2d98c945dd85d2d09ad0892670e2ba4f69e11b2acba2a876a3f6e
+size 927418280

special_tokens_map.json

@@ -0,0 +1,27 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": {
+    "content": "<|im_start|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcfe42da0a4497e8b2b172c1f9f4ec423a46dc12907f4349c55025f670422ba9
+size 11418266

tokenizer_config.json

@@ -0,0 +1,44 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": "<|im_start|>",
+  "chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.005887272119083459,
+    "total_flos": 0.0,
+    "train_loss": 4.74956651167436,
+    "train_runtime": 13.2482,
+    "train_samples_per_second": 24.154,
+    "train_steps_per_second": 1.51
+}

trainer_state.json

@@ -0,0 +1,102 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.005887272119083459,
+  "eval_steps": 400,
+  "global_step": 22,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001338016390700786,
+      "grad_norm": 1640.0,
+      "learning_rate": 7.464101615137755e-08,
+      "logits/chosen": -0.8070744276046753,
+      "logits/rejected": -0.6946424841880798,
+      "logps/chosen": -1.7985578775405884,
+      "logps/rejected": -1.9126713275909424,
+      "loss": 82.739,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": -17.985576629638672,
+      "rewards/margins": 1.141135334968567,
+      "rewards/rejected": -19.126712799072266,
+      "step": 5
+    },
+    {
+      "epoch": 0.002676032781401572,
+      "grad_norm": 1448.0,
+      "learning_rate": 4.6945927106677224e-08,
+      "logits/chosen": -0.6592462658882141,
+      "logits/rejected": -0.5781775712966919,
+      "logps/chosen": -1.62754225730896,
+      "logps/rejected": -1.859715461730957,
+      "loss": 82.6139,
+      "rewards/accuracies": 0.612500011920929,
+      "rewards/chosen": -16.27542495727539,
+      "rewards/margins": 2.3217310905456543,
+      "rewards/rejected": -18.59715461730957,
+      "step": 10
+    },
+    {
+      "epoch": 0.004014049172102358,
+      "grad_norm": 2352.0,
+      "learning_rate": 1.4288495612538426e-08,
+      "logits/chosen": -0.7022191882133484,
+      "logits/rejected": -0.6638501286506653,
+      "logps/chosen": -1.8131605386734009,
+      "logps/rejected": -1.7845131158828735,
+      "loss": 106.2027,
+      "rewards/accuracies": 0.5375000238418579,
+      "rewards/chosen": -18.131603240966797,
+      "rewards/margins": -0.2864739000797272,
+      "rewards/rejected": -17.84513282775879,
+      "step": 15
+    },
+    {
+      "epoch": 0.005352065562803144,
+      "grad_norm": 2496.0,
+      "learning_rate": 0.0,
+      "logits/chosen": -0.6888337135314941,
+      "logits/rejected": -0.650062084197998,
+      "logps/chosen": -1.786657691001892,
+      "logps/rejected": -1.9025501012802124,
+      "loss": 96.0509,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": -17.8665771484375,
+      "rewards/margins": 1.1589242219924927,
+      "rewards/rejected": -19.025501251220703,
+      "step": 20
+    },
+    {
+      "epoch": 0.005887272119083459,
+      "step": 22,
+      "total_flos": 0.0,
+      "train_loss": 4.74956651167436,
+      "train_runtime": 13.2482,
+      "train_samples_per_second": 24.154,
+      "train_steps_per_second": 1.51
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 20,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1000000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb6798d12fedf8f11141b2078040b1c5c1bcd5a298385f8634a32c2c5e6776a6
+size 5880