Set teacher ckptr output_dir to match student in KD configs (#2185)

pytorch · Dec 19, 2024 · 46a1ef0 · 46a1ef0
1 parent 0528c11
commit 46a1ef0
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 4 deletions.
diff --git a/recipes/configs/llama3_2/8B_to_1B_KD_lora_distributed.yaml b/recipes/configs/llama3_2/8B_to_1B_KD_lora_distributed.yaml
@@ -59,7 +59,7 @@ teacher_checkpointer:
     model-00004-of-00004.safetensors
   ]
   recipe_checkpoint: null
-  output_dir: /tmp/Meta-Llama-3.1-8B-Instruct/
+  output_dir: ${output_dir}
   model_type: LLAMA3
 
 # Dataset and Sampler

diff --git a/recipes/configs/llama3_2/8B_to_1B_KD_lora_single_device.yaml b/recipes/configs/llama3_2/8B_to_1B_KD_lora_single_device.yaml
@@ -59,7 +59,7 @@ teacher_checkpointer:
     model-00004-of-00004.safetensors
   ]
   recipe_checkpoint: null
-  output_dir: /tmp/Meta-Llama-3.1-8B-Instruct/
+  output_dir: ${output_dir}
   model_type: LLAMA3
 
 # Dataset and Sampler

diff --git a/recipes/configs/qwen2/1.5_to_0.5B_KD_lora_distributed.yaml b/recipes/configs/qwen2/1.5_to_0.5B_KD_lora_distributed.yaml
@@ -51,7 +51,7 @@ teacher_checkpointer:
     hf_model_0001_0.pt
   ]
   recipe_checkpoint: null
-  output_dir: /tmp/Qwen2-1.5B-Instruct-lora-finetune
+  output_dir: ${output_dir}
   model_type: QWEN2
 
 resume_from_checkpoint: False

diff --git a/recipes/configs/qwen2/1.5_to_0.5B_KD_lora_single_device.yaml b/recipes/configs/qwen2/1.5_to_0.5B_KD_lora_single_device.yaml
@@ -51,7 +51,7 @@ teacher_checkpointer:
     model.safetensors
   ]
   recipe_checkpoint: null
-  output_dir: /tmp/Qwen2-1.5B-Instruct
+  output_dir: ${output_dir}
   model_type: QWEN2
 
 resume_from_checkpoint: False