feat: Adds embed_tokens, lm_head as trainable for vocab expansion in peft and enables tying of adapters

pyproject.toml

@@ -35,7 +35,7 @@ dependencies = [
 "tokenizers<=0.22",
 "tqdm>=4.66.2,<5.0",
 "trl>=0.19.1,<0.20.0",
-"peft @ git+https://github.com/huggingface/peft.git@293aea5df6db240856a77f89955d1a89ce38b50d",
+"peft @ git+https://github.com/romitjain/peft.git@8388aa869473a60589a01e6950ea0583d3612783",
 "datasets>=4.0.0,<5.0.0",
 "simpleeval>=0.9.13,<2.0",
 "pillow>=11.0.0,<12.0",

tests/artifacts/language_models/__init__.py

@@ -20,3 +20,6 @@
 ### Constants used for model path
 PREDEFINED_MODEL_PATH = os.path.join(os.path.dirname(__file__))
 MAYKEYE_TINY_LLAMA_CACHED = os.path.join(PREDEFINED_MODEL_PATH, "maykeye-tinyllama-v0")
+TRL_INTERNAL_GEMMA_CACHED = os.path.join(
+    PREDEFINED_MODEL_PATH, "tiny-Gemma2ForCausalLM"
+)

tests/artifacts/language_models/tiny-Gemma2ForCausalLM/chat_template.jinja

@@ -0,0 +1,4 @@
+{{ bos_token }}{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '<start_of_turn>' + role + '
+' + message['content'] | trim + '<end_of_turn>
+' }}{% endfor %}{% if add_generation_prompt %}{{'<start_of_turn>model
+'}}{% endif %}

tests/artifacts/language_models/tiny-Gemma2ForCausalLM/config.json

@@ -0,0 +1,34 @@
+{
+  "architectures": [
+    "Gemma2ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "attn_logit_softcapping": 50.0,
+  "bos_token_id": 2,
+  "dtype": "float32",
+  "eos_token_id": 1,
+  "final_logit_softcapping": 30.0,
+  "head_dim": 256,
+  "hidden_activation": "gelu_pytorch_tanh",
+  "hidden_size": 8,
+  "initializer_range": 0.02,
+  "intermediate_size": 32,
+  "layer_types": [
+    "sliding_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 8192,
+  "model_type": "gemma2",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 2,
+  "num_key_value_heads": 2,
+  "pad_token_id": 0,
+  "query_pre_attn_scalar": 256,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "transformers_version": "4.57.0.dev0",
+  "use_cache": true,
+  "vocab_size": 256000
+}

tests/artifacts/language_models/tiny-Gemma2ForCausalLM/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 2,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.57.0.dev0"
+}

tests/artifacts/language_models/tiny-Gemma2ForCausalLM/special_tokens_map.json

@@ -0,0 +1,34 @@
+{
+  "additional_special_tokens": [
+    "<start_of_turn>",
+    "<end_of_turn>"
+  ],
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}