update

.github/scripts/oc_score_assert.py

@@ -7,18 +7,22 @@
 output_path = 'regression_result_daily'
 
 chat_model_list = [
-    'baichuan2-7b-chat-hf', 'deepseek-7b-chat-hf', 'deepseek-moe-16b-chat-hf',
-    'deepseek-7b-chat-vllm', 'internlm2_5-7b-chat-hf',
-    'internlm2_5-7b-chat-turbomind', 'internlm2-chat-1.8b-turbomind',
+    'baichuan2-7b-chat-hf', 'glm-4-9b-chat-turbomind', 'glm-4-9b-chat-vllm',
+    'deepseek-7b-chat-hf', 'deepseek-moe-16b-chat-hf',
+    'deepseek-v2-lite-chat-hf', 'deepseek-7b-chat-vllm', 'gemma2-2b-it-hf',
+    'gemma2-9b-it-hf', 'gemma-7b-it-vllm', 'internlm2_5-7b-chat-hf',
+    'internlm2_5-20b-chat-hf', 'internlm2_5-7b-chat-turbomind',
+    'internlm2_5-20b-chat-turbomind', 'internlm2-chat-1.8b-turbomind',
     'internlm2-chat-1.8b-sft-turbomind', 'internlm2-chat-7b-turbomind',
     'internlm2-chat-7b-sft-turbomind', 'internlm2-chat-7b-vllm',
-    'llama-3-8b-instruct-hf', 'llama-3-8b-instruct-turbomind',
-    'mistral-7b-instruct-v0.2-hf', 'mistral-7b-instruct-v0.2-vllm',
+    'llama-3_1-8b-instruct-hf', 'llama-3-8b-instruct-hf',
+    'llama-3_1-8b-instruct-turbomind', 'llama-3-8b-instruct-turbomind',
+    'mistral-7b-instruct-v0.3-hf', 'mistral-7b-instruct-v0.2-vllm',
     'minicpm-2b-dpo-fp32-hf', 'minicpm-2b-sft-bf16-hf',
     'minicpm-2b-sft-fp32-hf', 'phi-3-mini-4k-instruct-hf',
-    'qwen1.5-0.5b-chat-hf', 'qwen2-1.5b-instruct-turbomind',
-    'qwen2-7b-instruct-turbomind', 'qwen1.5-0.5b-chat-vllm',
-    'yi-1.5-6b-chat-hf', 'yi-1.5-9b-chat-hf', 'lmdeploy-api-test'
+    'qwen1.5-0.5b-chat-hf', 'qwen2-1.5b-instruct-hf', 'qwen2-7b-instruct-hf',
+    'qwen2-1.5b-instruct-turbomind', 'qwen2-7b-instruct-turbomind',
+    'qwen1.5-0.5b-chat-vllm', 'yi-1.5-6b-chat-hf', 'yi-1.5-9b-chat-hf'
 ]
 base_model_list = [
     'deepseek-moe-16b-base-hf', 'deepseek-7b-base-turbomind',

.github/scripts/oc_score_baseline.yaml

@@ -8,6 +8,16 @@ glm-4-9b-chat-hf:
     race-middle: 88
     race-high: 88
 
+glm-4-9b-chat-turbomind:
+    gsm8k: 69
+    race-middle: 82
+    race-high: 77
+
+glm-4-9b-chat-vllm:
+    gsm8k: 73
+    race-middle: 87
+    race-high: 87
+
 deepseek-7b-chat-hf:
     gsm8k: 60
     race-middle: 74
@@ -18,6 +28,11 @@ deepseek-moe-16b-chat-hf:
     race-middle: 62
     race-high: 70
 
+deepseek-v2-lite-chat-hf:
+    gsm8k: 59
+    race-middle: 82
+    race-high: 79
+
 deepseek-7b-chat-vllm:
     gsm8k: 63
     race-middle: 74
@@ -33,16 +48,36 @@ gemma-7b-it-hf:
     race-middle: 74
     race-high: 71
 
+gemma2-2b-it-hf:
+    gsm8k: 62
+    race-middle: 75
+    race-high: 67
+
+gemma2-9b-it-hf:
+    gsm8k: 80
+    race-middle: 89
+    race-high: 85
+
 internlm2_5-7b-chat-hf:
     gsm8k: 86
     race-middle: 92
     race-high: 93
 
+internlm2_5-20b-chat-hf:
+    gsm8k: 91
+    race-middle: 95
+    race-high: 91
+
 internlm2_5-7b-chat-turbomind:
     gsm8k: 87
     race-middle: 92
     race-high: 93
 
+internlm2_5-20b-chat-turbomind:
+    gsm8k: 91
+    race-middle: 95
+    race-high: 91
+
 internlm2-chat-1.8b-turbomind:
     gsm8k: 40
     race-middle: 82
@@ -68,11 +103,21 @@ internlm2-chat-7b-vllm:
     race-middle: 90
     race-high: 91
 
+llama-3_1-8b-instruct-hf:
+    gsm8k: 82
+    race-middle: 82
+    race-high: 88
+
 llama-3-8b-instruct-hf:
     gsm8k: 77
     race-middle: 85
     race-high: 87
 
+llama-3_1-8b-instruct-turbomind:
+    gsm8k: 79
+    race-middle: 82
+    race-high: 88
+
 llama-3-8b-instruct-turbomind:
     gsm8k: 77
     race-middle: 85
@@ -83,6 +128,11 @@ mistral-7b-instruct-v0.2-hf:
     race-middle: 82
     race-high: 78
 
+mistral-7b-instruct-v0.3-hf:
+    gsm8k: 53
+    race-middle: 80
+    race-high: 78
+
 mistral-7b-instruct-v0.2-vllm:
     gsm8k: 49
     race-middle: 81
@@ -118,6 +168,11 @@ qwen1.5-0.5b-chat-hf:
     race-middle: 55
     race-high: 50
 
+qwen2-1.5b-instruct-hf:
+    gsm8k: 63
+    race-middle: 77
+    race-high: 86
+
 qwen2-1.5b-instruct-turbomind:
     gsm8k: 60
     race-middle: 77
@@ -128,6 +183,11 @@ qwen2-7b-instruct-turbomind:
     race-middle: 87
     race-high: 89
 
+qwen2-7b-instruct-hf:
+    gsm8k: 85
+    race-middle: 87
+    race-high: 91
+
 qwen1.5-0.5b-chat-vllm:
     gsm8k: 5
     race-middle: 57