Fix VLLM argument error (#1207)

configs/models/mistral/vllm_mistral_7b_instruct_v0_1.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=1, num_procs=1),
     )
 ]

configs/models/mistral/vllm_mistral_7b_instruct_v0_2.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=1, num_procs=1),
     )
 ]

configs/models/mistral/vllm_mixtral_8x7b_instruct_v0_1.py

@@ -20,7 +20,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=2, num_procs=1),
     )
 ]

configs/models/others/vllm_orionstar_14b_longchat.py

@@ -21,6 +21,6 @@ models = [
         max_seq_len=4096,
         batch_size=32,
         run_cfg=dict(num_gpus=4, num_procs=1),
-        end_str='<|endoftext|>',
+        stop_words=['<|endoftext|>'],
     )
 ]

configs/models/qwen/vllm_qwen1_5_14b_chat.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='<|im_end|>',
+        stop_words=['<|im_end|>'],
         run_cfg=dict(num_gpus=2, num_procs=1),
     )
 ]

configs/models/qwen/vllm_qwen1_5_72b_chat.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='<|im_end|>',
+        stop_words=['<|im_end|>'],
         run_cfg=dict(num_gpus=4, num_procs=1),
     )
 ]

configs/models/qwen/vllm_qwen_14b_chat.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='<|im_end|>',
+        stop_words=['<|im_end|>'],
         run_cfg=dict(num_gpus=4, num_procs=1),
     )
 ]

configs/models/qwen/vllm_qwen_72b_chat.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='<|im_end|>',
+        stop_words=['<|im_end|>'],
         run_cfg=dict(num_gpus=4, num_procs=1),
     )
 ]

configs/models/vicuna/vllm_vicuna_13b_v15_16k.py

@@ -17,7 +17,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=2, num_procs=1),
     )
 ]

configs/models/vicuna/vllm_vicuna_7b_v15_16k.py

@@ -17,7 +17,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=1, num_procs=1),
     )
 ]

configs/models/wizardlm/vllm_wizardlm_13b_v1_2.py

@@ -18,7 +18,7 @@ models = [
         max_seq_len=2048,
         batch_size=1,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=1, num_procs=1),
     )
 ]

configs/models/wizardlm/vllm_wizardlm_70b_v1_0.py

@@ -19,7 +19,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=4, num_procs=1),
     )
 ]

configs/models/wizardlm/vllm_wizardlm_7b_v1_0.py

@@ -18,7 +18,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=1, num_procs=1),
     )
 ]

configs/models/zephyr/vllm_zephyr_7b_beta.py

@@ -17,7 +17,7 @@ models = [
         max_seq_len=2048,
         batch_size=32,
         generation_kwargs=dict(temperature=0),
-        end_str='</s>',
+        stop_words=['</s>'],
         run_cfg=dict(num_gpus=1, num_procs=1),
     )
 ]