From 7327e3a4676402a87d1a0779759bc2a788279529 Mon Sep 17 00:00:00 2001 From: linjunyao Date: Tue, 7 Jan 2025 08:49:57 +0000 Subject: [PATCH] fixed typo for mistral model configs --- ...> lmdeploy_mistral_large_instruct_2407.py} | 6 ++---- .../lmdeploy_mistral_large_instruct_2411.py | 20 +++++++++++++++++++ .../lmdeploy_mistral_nemo_instruct_2407.py | 15 ++++++++++++++ .../lmdeploy_mistral_small_instruct_2409.py | 15 ++++++++++++++ ...> lmdeploy_mistral_large_instruct_2407.py} | 6 ++---- .../lmdeploy_mistral_large_instruct_2411.py | 6 ++---- .../lmdeploy_mistral_small_instruct_2409.py | 4 ++-- 7 files changed, 58 insertions(+), 14 deletions(-) rename configs/models/mistral/{lmdeploy_mixtral_large_instruct_2407.py => lmdeploy_mistral_large_instruct_2407.py} (74%) create mode 100644 configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py create mode 100644 configs/models/mistral/lmdeploy_mistral_nemo_instruct_2407.py create mode 100644 configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py rename opencompass/configs/models/mistral/{lmdeploy_mixtral_large_instruct_2407.py => lmdeploy_mistral_large_instruct_2407.py} (74%) diff --git a/configs/models/mistral/lmdeploy_mixtral_large_instruct_2407.py b/configs/models/mistral/lmdeploy_mistral_large_instruct_2407.py similarity index 74% rename from configs/models/mistral/lmdeploy_mixtral_large_instruct_2407.py rename to configs/models/mistral/lmdeploy_mistral_large_instruct_2407.py index 6dcdce83..02b65e53 100644 --- a/configs/models/mistral/lmdeploy_mixtral_large_instruct_2407.py +++ b/configs/models/mistral/lmdeploy_mistral_large_instruct_2407.py @@ -3,7 +3,7 @@ from opencompass.models import TurboMindModelwithChatTemplate models = [ dict( type=TurboMindModelwithChatTemplate, - abbr='mixtral-large-instruct-2407-turbomind', + abbr='mistral-large-instruct-2407-turbomind', path='mistralai/Mistral-Large-Instruct-2407', engine_config=dict( session_len=32768, @@ -11,9 +11,7 @@ models = [ tp=4, cache_max_entry_count=0.7, ), - gen_config=dict( - top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096 - ), + gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), max_seq_len=32768, max_out_len=4096, batch_size=8, diff --git a/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py b/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py new file mode 100644 index 00000000..4e7fb0c7 --- /dev/null +++ b/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py @@ -0,0 +1,20 @@ +from opencompass.models import TurboMindModelwithChatTemplate + +models = [ + dict( + type=TurboMindModelwithChatTemplate, + abbr='mistral-large-instruct-2411-turbomind', + path='mistralai/Mistral-Large-Instruct-2411', + engine_config=dict( + session_len=32768, + max_batch_size=16, + tp=4, + cache_max_entry_count=0.7, + ), + gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), + max_seq_len=32768, + max_out_len=4096, + batch_size=8, + run_cfg=dict(num_gpus=4), + ) +] diff --git a/configs/models/mistral/lmdeploy_mistral_nemo_instruct_2407.py b/configs/models/mistral/lmdeploy_mistral_nemo_instruct_2407.py new file mode 100644 index 00000000..5e3c27f4 --- /dev/null +++ b/configs/models/mistral/lmdeploy_mistral_nemo_instruct_2407.py @@ -0,0 +1,15 @@ +from opencompass.models import TurboMindModelwithChatTemplate + +models = [ + dict( + type=TurboMindModelwithChatTemplate, + abbr='mistral-nemo-instruct-2407-turbomind', + path='mistralai/Mistral-Nemo-Instruct-2407', + engine_config=dict(session_len=32768, max_batch_size=16, tp=1), + gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), + max_seq_len=32768, + max_out_len=4096, + batch_size=16, + run_cfg=dict(num_gpus=1), + ) +] diff --git a/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py b/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py new file mode 100644 index 00000000..b6c07212 --- /dev/null +++ b/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py @@ -0,0 +1,15 @@ +from opencompass.models import TurboMindModelwithChatTemplate + +models = [ + dict( + type=TurboMindModelwithChatTemplate, + abbr='mistral-small-instruct-2409-turbomind', + path='mistralai/Mistral-Small-Instruct-2409', + engine_config=dict(session_len=32768, max_batch_size=16, tp=2), + gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), + max_seq_len=32768, + max_out_len=4096, + batch_size=16, + run_cfg=dict(num_gpus=2), + ) +] diff --git a/opencompass/configs/models/mistral/lmdeploy_mixtral_large_instruct_2407.py b/opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2407.py similarity index 74% rename from opencompass/configs/models/mistral/lmdeploy_mixtral_large_instruct_2407.py rename to opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2407.py index 6dcdce83..02b65e53 100644 --- a/opencompass/configs/models/mistral/lmdeploy_mixtral_large_instruct_2407.py +++ b/opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2407.py @@ -3,7 +3,7 @@ from opencompass.models import TurboMindModelwithChatTemplate models = [ dict( type=TurboMindModelwithChatTemplate, - abbr='mixtral-large-instruct-2407-turbomind', + abbr='mistral-large-instruct-2407-turbomind', path='mistralai/Mistral-Large-Instruct-2407', engine_config=dict( session_len=32768, @@ -11,9 +11,7 @@ models = [ tp=4, cache_max_entry_count=0.7, ), - gen_config=dict( - top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096 - ), + gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), max_seq_len=32768, max_out_len=4096, batch_size=8, diff --git a/opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py b/opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py index 205dc27b..4e7fb0c7 100644 --- a/opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py +++ b/opencompass/configs/models/mistral/lmdeploy_mistral_large_instruct_2411.py @@ -3,7 +3,7 @@ from opencompass.models import TurboMindModelwithChatTemplate models = [ dict( type=TurboMindModelwithChatTemplate, - abbr='mixtral-large-instruct-2411-turbomind', + abbr='mistral-large-instruct-2411-turbomind', path='mistralai/Mistral-Large-Instruct-2411', engine_config=dict( session_len=32768, @@ -11,9 +11,7 @@ models = [ tp=4, cache_max_entry_count=0.7, ), - gen_config=dict( - top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096 - ), + gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), max_seq_len=32768, max_out_len=4096, batch_size=8, diff --git a/opencompass/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py b/opencompass/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py index 1b5ac010..b6c07212 100644 --- a/opencompass/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py +++ b/opencompass/configs/models/mistral/lmdeploy_mistral_small_instruct_2409.py @@ -3,8 +3,8 @@ from opencompass.models import TurboMindModelwithChatTemplate models = [ dict( type=TurboMindModelwithChatTemplate, - abbr="mistral-small-instruct-2409-turbomind", - path="mistralai/Mistral-Small-Instruct-2409", + abbr='mistral-small-instruct-2409-turbomind', + path='mistralai/Mistral-Small-Instruct-2409', engine_config=dict(session_len=32768, max_batch_size=16, tp=2), gen_config=dict(top_k=1, temperature=1e-6, top_p=0.9, max_new_tokens=4096), max_seq_len=32768,