Updated HF Link in model_configs for Llama-3-8B-Instruct-SkillMix (#409)

* updated link to hf * updated link to hf --------- Co-authored-by: Simon Park <[email protected]>
tatsu-lab · Sep 20, 2024 · f8a7bf9 · f8a7bf9
1 parent 57c3fa0
commit f8a7bf9
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/src/alpaca_eval/models_configs/Llama-3-8B-Instruct-SkillMix/configs.yaml b/src/alpaca_eval/models_configs/Llama-3-8B-Instruct-SkillMix/configs.yaml
@@ -2,7 +2,7 @@ Llama-3-8B-Instruct-SkillMix:
   prompt_template: "alpaca-7b/prompt.txt"
   fn_completions: "vllm_local_completions"
   completions_kwargs:
-    model_name: "/scratch/gpfs/PLI/alpaca_synth/checkpoints_ft_torchtune/llama3_8b_base_k2_4000_examples_diverse_dataset_long_instructions_rebuttals/opt_adamw/lr_0.00002/bs_128/epoch_7"
+    model_name: "PrincetonPLI/Llama-3-8B-Instruct-SkillMix"
     batch_size: 900
     model_kwargs:
       torch_dtype: 'bfloat16'
@@ -14,6 +14,6 @@ Llama-3-8B-Instruct-SkillMix:
     - 128001
     - 128009
   pretty_name: "Llama-3-8B-Instruct-SkillMix"
-  link: "https://huggingface.co/meta-llama/Meta-Llama-3-8B"
+  link: "https://huggingface.co/PrincetonPLI/Llama-3-8B-Instruct-SkillMix"