Checkpoint after epoch 160

README.md

@@ -0,0 +1,44 @@
+---
+license: apache-2.0
+library_name: transformers
+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="ajagota71//kaggle/working/irl_llms/outputs/2025-05-16_09-29-11/checkpoints/temp-checkpoint-epoch-160")
+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("ajagota71//kaggle/working/irl_llms/outputs/2025-05-16_09-29-11/checkpoints/temp-checkpoint-epoch-160")
+model = AutoModelForCausalLMWithValueHead.from_pretrained("ajagota71//kaggle/working/irl_llms/outputs/2025-05-16_09-29-11/checkpoints/temp-checkpoint-epoch-160")
+
+inputs = tokenizer("Hello, my llama is cute", return_tensors="pt")
+outputs = model(**inputs, labels=inputs["input_ids"])
+```

config.json

@@ -0,0 +1,31 @@
+{
+  "architectures": [
+    "GPTNeoXForCausalLM"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "classifier_dropout": 0.1,
+  "eos_token_id": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 512,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
+  "num_attention_heads": 8,
+  "num_hidden_layers": 6,
+  "partial_rotary_factor": 0.25,
+  "rope_scaling": null,
+  "rope_theta": 10000,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.25,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_parallel_residual": true,
+  "vocab_size": 50304
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "transformers_version": "4.51.3"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b32be9d1a669852be11745230be30816157c2f3cc2c7c5b1f5e4d1837d9def3e
+size 281717404

rlhf_config.yaml

@@ -0,0 +1,149 @@
+rlhf:
+  model:
+    name: EleutherAI/pythia-70M
+    learning_rate: 1.0e-05
+    ppo_epochs: 4
+    init_kl_coef: 0.2
+    target: 6
+    cliprange: 0.2
+    cliprange_value: 0.2
+    vf_coef: 0.1
+    adap_kl_ctrl: true
+    use_score_norm: true
+    ratio_threshold: 10.0
+    batch_size: 128
+    mini_batch_size: 8
+    forward_batch_size: 8
+    gradient_accumulation_steps: 8
+    reward_model: facebook/roberta-hate-speech-dynabench-r4-target
+    use_raw_logits: true
+    generation:
+      min_length: 5
+      output_min_length: 15
+      output_max_length: 20
+      do_sample: true
+      top_k: 0.0
+      top_p: 1.0
+  training:
+    num_train_epochs: 100
+    save_freq: 20
+    eval_freq: 20
+    seed: 42
+  dataset:
+    name: allenai/real-toxicity-prompts
+    toxicity_threshold: 0.3
+    input_min_text_length: 15
+    input_max_text_length: 20
+    test_size: 0.1
+  output:
+    push_to_hub: true
+    organization: null
+    repository_name: pythia-70m-detox
+  wandb:
+    project: irl_llms
+    entity: null
+    name: null
+irl:
+  mode: train
+  dataset:
+    original_model_name: EleutherAI/pythia-70M
+    detoxified_model_name: ajagota71/pythia-70m-detox-epoch-100
+    original_dataset_path: null
+    detoxified_dataset_path: null
+    cache_dir: ${hydra:runtime.cwd}/datasets
+    num_samples: 1000
+    max_new_tokens: 30
+    batch_size: 16
+    temperature: 0.7
+    top_p: 1.0
+    seed: ${seed}
+    use_cached: false
+    toxicity_threshold: 0.3
+    push_to_hub: false
+    hub_org: null
+    hub_token: ${oc.env:HF_TOKEN,null}
+    private: false
+    use_half_precision: null
+  model:
+    reward_model_base: null
+    use_half_precision: null
+    num_unfrozen_layers: 1
+  training:
+    irl_method: max_margin
+    learning_rate: 1.0e-05
+    epochs: 20
+    batch_size: 4
+    eval_interval: 5
+    max_length: 512
+    train_test_split: 0.8
+    grad_clip: 1.0
+    weight_decay: 0.01
+    margin: 0.1
+    temperature: 0.1
+    adam_epsilon: 1.0e-08
+    seed: ${seed}
+    include_prompt: true
+  output:
+    repo_name_prefix: irl-reward-model
+    base_dir: ${hydra:runtime.cwd}/outputs/irl
+    save_checkpoints: true
+    push_to_hub: false
+    hub_org: ajagota71
+    private: false
+  evaluation:
+    true_reward_model: facebook/roberta-hate-speech-dynabench-r4-target
+  logging:
+    project_name: irl-detoxification
+    use_wandb: true
+    wandb_mode: online
+now: 2025-05-16_09-29-11
+seed: 42
+output_dir: ${hydra:runtime.cwd}/outputs/${now:%Y-%m-%d_%H-%M-%S}
+mode: train
+model:
+  name: EleutherAI/pythia-70M
+  learning_rate: 1.0e-05
+  batch_size: 128
+  mini_batch_size: 8
+  forward_batch_size: 8
+  gradient_accumulation_steps: 8
+  reward_model: facebook/roberta-hate-speech-dynabench-r4-target
+  use_raw_logits: true
+  ppo_epochs: 4
+  init_kl_coef: 0.2
+  target: 6
+  cliprange: 0.2
+  cliprange_value: 0.2
+  vf_coef: 0.1
+  adap_kl_ctrl: true
+  use_score_norm: true
+  ratio_threshold: 10.0
+  generation:
+    min_length: 5
+    output_min_length: 15
+    output_max_length: 20
+    do_sample: true
+    top_k: 0.0
+    top_p: 1.0
+training:
+  num_train_epochs: 200
+  save_freq: 50
+  eval_freq: 20
+  seed: 42
+dataset:
+  name: allenai/real-toxicity-prompts
+  toxicity_threshold: 0.3
+  input_min_text_length: 15
+  input_max_text_length: 20
+  test_size: 0.1
+output:
+  push_to_hub: true
+  push_checkpoints_to_hub: true
+  checkpoint_push_freq: 20
+  organization: ajagota71
+  repository_name: pythia-70m-fb-detox
+  private: false
+wandb:
+  project: irl_llms
+  entity: null
+  name: pythia-70M-2025-05-16_09-29-11

special_tokens_map.json

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

tokenizer_config.json

@@ -0,0 +1,215 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<|padding|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50254": {
+      "content": "                        ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50255": {
+      "content": "                       ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50256": {
+      "content": "                      ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50257": {
+      "content": "                     ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50258": {
+      "content": "                    ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50259": {
+      "content": "                   ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50260": {
+      "content": "                  ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50261": {
+      "content": "                 ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50262": {
+      "content": "                ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50263": {
+      "content": "               ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50264": {
+      "content": "              ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50265": {
+      "content": "             ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50266": {
+      "content": "            ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50267": {
+      "content": "           ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50268": {
+      "content": "          ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50269": {
+      "content": "         ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50270": {
+      "content": "        ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50271": {
+      "content": "       ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50272": {
+      "content": "      ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50273": {
+      "content": "     ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50274": {
+      "content": "    ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50275": {
+      "content": "   ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50276": {
+      "content": "  ",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "GPTNeoXTokenizer",
+  "unk_token": "<|endoftext|>"
+}