Add SelfMoA_gemma-2-9b-it-SimPO, SelfMoA_gemma-2-9b-it-WPO-HB to Alpa…

…caEval
tatsu-lab · Sep 24, 2024 · e982df7 · e982df7
1 parent f8a7bf9
commit e982df7
Show file tree

Hide file tree

Showing 8 changed files with 172,380 additions and 210 deletions.
diff --git a/results/SelfMoA_gemma-2-9b-it-SimPO/model_outputs.json b/results/SelfMoA_gemma-2-9b-it-SimPO/model_outputs.json
diff --git a/results/SelfMoA_gemma-2-9b-it-SimPO/weighted_alpaca_eval_gpt4_turbo/annotations.json b/results/SelfMoA_gemma-2-9b-it-SimPO/weighted_alpaca_eval_gpt4_turbo/annotations.json
diff --git a/results/SelfMoA_gemma-2-9b-it-WPO-HB/model_outputs.json b/results/SelfMoA_gemma-2-9b-it-WPO-HB/model_outputs.json
diff --git a/results/SelfMoA_gemma-2-9b-it-WPO-HB/weighted_alpaca_eval_gpt4_turbo/annotations.json b/results/SelfMoA_gemma-2-9b-it-WPO-HB/weighted_alpaca_eval_gpt4_turbo/annotations.json
diff --git a/...lpaca_eval/leaderboards/data_AlpacaEval_2/weighted_alpaca_eval_gpt4_turbo_leaderboard.csv b/...lpaca_eval/leaderboards/data_AlpacaEval_2/weighted_alpaca_eval_gpt4_turbo_leaderboard.csv
diff --git a/...ights/weighted_alpaca_eval_gpt4_turbo/length_controlled_v1/baseline_gpt4_1106_preview.csv b/...ights/weighted_alpaca_eval_gpt4_turbo/length_controlled_v1/baseline_gpt4_1106_preview.csv
@@ -187,3 +187,5 @@ Shopee-SlimMoA-v1,-0.6930943742294789,0.5778443790027642,1.4506276222723822
 blendaxai-gm-l6-vo31,-1.4827230167114802,0.8256378421072179,1.5942312525409852
 Llama-3-8B-Instruct-SkillMix,-0.3007600604906024,0.4853486472189554,-0.2808727525336005
 REBEL-Llama-3-8B-Instruct-Armo,-1.0427168605260002,0.6464073051877255,0.0395191056877229
+SelfMoA_gemma-2-9b-it-SimPO,-0.8425253084188749,0.5482697859900880,1.2874783673834935
+SelfMoA_gemma-2-9b-it-WPO-HB,0.2523363342614252,0.3970191588440620,1.4137351138484051
diff --git a/src/alpaca_eval/models_configs/SelfMoA_gemma-2-9b-it-SimPO/configs.yaml b/src/alpaca_eval/models_configs/SelfMoA_gemma-2-9b-it-SimPO/configs.yaml
@@ -0,0 +1,9 @@
+SelfMoA_gemma-2-9b-it-SimPO:
+  prompt_template: "gemma-2-9b-it-DPO/prompt.txt"
+  fn_completions: null
+  completions_kwargs:
+    max_new_tokens: 2048
+    model_name: "SelfMoA_gemma-2-9b-it-SimPO"
+    temperature: 0.7
+  pretty_name: "SelfMoA + gemma-2-9b-it-SimPO"
+  link: "https://github.com/wenzhe-li/Self-MoA/"
diff --git a/src/alpaca_eval/models_configs/SelfMoA_gemma-2-9b-it-WPO-HB/configs.yaml b/src/alpaca_eval/models_configs/SelfMoA_gemma-2-9b-it-WPO-HB/configs.yaml
@@ -0,0 +1,9 @@
+SelfMoA_gemma-2-9b-it-WPO-HB:
+  prompt_template: "gemma-2-9b-it-DPO/prompt.txt"
+  fn_completions: null
+  completions_kwargs:
+    max_new_tokens: 2048
+    model_name: "SelfMoA_gemma-2-9b-it-WPO-HB"
+    temperature: 0.7
+  pretty_name: "SelfMoA + gemma-2-9b-it-WPO-HB"
+  link: "https://github.com/wenzhe-li/Self-MoA/"