support for QAT w RL (DPO) (#2776)

2025-06-13 10:00:35 -04:00
parent eac4a61f55
commit b2274d430b
9 changed files with 152 additions and 32 deletions
--- a/examples/llama-3/instruct-dpo-lora-8b.yml
+++ b/examples/llama-3/instruct-dpo-lora-8b.yml
@@ -5,6 +5,10 @@ tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name

+special_tokens:
+  pad_token: <|finetune_right_pad_id|>
+  eos_token: <|eot_id|>
+
 load_in_8bit: true
 load_in_4bit: false