Step 100 MA=0.712

README.md

@@ -0,0 +1,43 @@
+---
+license: apache-2.0
+tags:
+- trl
+- ppo
+- transformers
+- reinforcement-learning
+---
+
+# TRL Model
+
+This is a [TRL language model](https://github.com/huggingface/trl) that has been fine-tuned with reinforcement learning to
+ guide the model outputs according to a value, function, or human feedback. The model can be used for text generation.
+
+## Usage
+
+To use this model for inference, first install the TRL library:
+
+```bash
+python -m pip install trl
+```
+
+You can then generate text as follows:
+
+```python
+from transformers import pipeline
+
+generator = pipeline("text-generation", model="MeowFR//workspace/grpo-mistral-desklib/step-100")
+outputs = generator("Hello, my llama is cute")
+```
+
+If you want to use the model for training or to obtain the outputs from the value head, load the model as follows:
+
+```python
+from transformers import AutoTokenizer
+from trl import AutoModelForCausalLMWithValueHead
+
+tokenizer = AutoTokenizer.from_pretrained("MeowFR//workspace/grpo-mistral-desklib/step-100")
+model = AutoModelForCausalLMWithValueHead.from_pretrained("MeowFR//workspace/grpo-mistral-desklib/step-100")
+
+inputs = tokenizer("Hello, my llama is cute", return_tensors="pt")
+outputs = model(**inputs, labels=inputs["input_ids"])
+```

adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "mistralai/Mistral-7B-Instruct-v0.2",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "o_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f0529bc23aef398d7307fa041572482cb6e95f2067f597c232e46145ac014fa
+size 54560368

config.json

@@ -0,0 +1,50 @@
+{
+  "accelerator_kwargs": {},
+  "adap_kl_ctrl": true,
+  "backward_batch_size": 2,
+  "batch_size": 4,
+  "cliprange": 0.2,
+  "cliprange_value": 0.2,
+  "compare_steps": 1,
+  "early_stopping": false,
+  "exp_name": "ipykernel_launcher",
+  "forward_batch_size": null,
+  "gamma": 1,
+  "global_backward_batch_size": 2,
+  "global_batch_size": 4,
+  "gradient_accumulation_steps": 2,
+  "gradient_checkpointing": false,
+  "horizon": 10000,
+  "init_kl_coef": 0.05,
+  "is_encoder_decoder": false,
+  "is_peft_model": true,
+  "kl_penalty": "kl",
+  "lam": 0.95,
+  "learning_rate": 6.5e-06,
+  "log_with": null,
+  "max_grad_norm": null,
+  "mini_batch_size": 1,
+  "model_name": "gpt2",
+  "optimize_cuda_cache": null,
+  "optimize_device_cache": false,
+  "ppo_epochs": 4,
+  "project_kwargs": {},
+  "push_to_hub_if_best_kwargs": {},
+  "query_dataset": "imdb",
+  "ratio_threshold": 10.0,
+  "remove_unused_columns": true,
+  "reward_model": "sentiment-analysis:lvwerra/distilbert-imdb",
+  "score_clip": null,
+  "seed": 0,
+  "steps": 20000,
+  "target": 6,
+  "target_kl": 0.5,
+  "task_name": null,
+  "tracker_kwargs": {},
+  "tracker_project_name": "trl",
+  "use_score_norm": false,
+  "use_score_scaling": false,
+  "vf_coef": 0.1,
+  "whiten_rewards": false,
+  "world_size": 1
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55dc21cd02ebfc9f708776601147fb4b9fc4e79de23cd51e479a88258832d711
+size 17916

special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443

tokenizer_config.json

@@ -0,0 +1,44 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "chat_template": "{%- if messages[0]['role'] == 'system' %}\n    {%- set system_message = messages[0]['content'] %}\n    {%- set loop_messages = messages[1:] %}\n{%- else %}\n    {%- set loop_messages = messages %}\n{%- endif %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n        {{- raise_exception('After the optional system message, conversation roles must alternate user/assistant/user/assistant/...') }}\n    {%- endif %}\n    {%- if message['role'] == 'user' %}\n        {%- if loop.first and system_message is defined %}\n            {{- ' [INST] ' + system_message + '\\n\\n' + message['content'] + ' [/INST]' }}\n        {%- else %}\n            {{- ' [INST] ' + message['content'] + ' [/INST]' }}\n        {%- endif %}\n    {%- elif message['role'] == 'assistant' %}\n        {{- ' ' + message['content'] + eos_token}}\n    {%- else %}\n        {{- raise_exception('Only user and assistant roles are supported, with the exception of an initial optional system message!') }}\n    {%- endif %}\n{%- endfor %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}