diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/deepdeek-coder-v2-instruct.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/deepdeek-coder-v2-instruct.yaml new file mode 100644 index 0000000000..d4431179e5 --- /dev/null +++ b/api/core/model_runtime/model_providers/siliconflow/llm/deepdeek-coder-v2-instruct.yaml @@ -0,0 +1,30 @@ +model: deepseek-ai/DeepSeek-Coder-V2-Instruct +label: + en_US: deepseek-ai/DeepSeek-Coder-V2-Instruct +model_type: llm +features: + - agent-thought +model_properties: + mode: chat + context_size: 32768 +parameter_rules: + - name: temperature + use_template: temperature + - name: max_tokens + use_template: max_tokens + type: int + default: 512 + min: 1 + max: 4096 + help: + zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。 + en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter. + - name: top_p + use_template: top_p + - name: frequency_penalty + use_template: frequency_penalty +pricing: + input: '1.33' + output: '1.33' + unit: '0.000001' + currency: RMB diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/deepseek-v2-chat.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/deepseek-v2-chat.yaml index da58e822f9..3926568db6 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/deepseek-v2-chat.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/deepseek-v2-chat.yaml @@ -1,11 +1,9 @@ model: deepseek-ai/deepseek-v2-chat label: - en_US: deepseek-ai/deepseek-v2-chat + en_US: deepseek-ai/DeepSeek-V2-Chat model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 32768 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/glm4-9b-chat.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/glm4-9b-chat.yaml index 115fc50b94..d6a4b21b66 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/glm4-9b-chat.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/glm4-9b-chat.yaml @@ -1,11 +1,9 @@ model: zhipuai/glm4-9B-chat label: - en_US: zhipuai/glm4-9B-chat + en_US: THUDM/glm-4-9b-chat model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 32768 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-57b-a14b-instruct.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-57b-a14b-instruct.yaml index 75eca7720c..39624dc5b9 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-57b-a14b-instruct.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-57b-a14b-instruct.yaml @@ -1,11 +1,9 @@ model: alibaba/Qwen2-57B-A14B-Instruct label: - en_US: alibaba/Qwen2-57B-A14B-Instruct + en_US: Qwen/Qwen2-57B-A14B-Instruct model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 32768 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-72b-instruct.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-72b-instruct.yaml index fcbc9e0b68..fb7ff6cb14 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-72b-instruct.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-72b-instruct.yaml @@ -1,11 +1,9 @@ model: alibaba/Qwen2-72B-Instruct label: - en_US: alibaba/Qwen2-72B-Instruct + en_US: Qwen/Qwen2-72B-Instruct model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 32768 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-7b-instruct.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-7b-instruct.yaml index eda1d40642..efda4abbd9 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-7b-instruct.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/qwen2-7b-instruct.yaml @@ -1,11 +1,9 @@ model: alibaba/Qwen2-7B-Instruct label: - en_US: alibaba/Qwen2-7B-Instruct + en_US: Qwen/Qwen2-7B-Instruct model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 32768 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-34b-chat.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-34b-chat.yaml index 6656e663e9..864ba46f1a 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-34b-chat.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-34b-chat.yaml @@ -1,11 +1,9 @@ model: 01-ai/Yi-1.5-34B-Chat label: - en_US: 01-ai/Yi-1.5-34B-Chat + en_US: 01-ai/Yi-1.5-34B-Chat-16K model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 16384 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-6b-chat.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-6b-chat.yaml index ba6e0c5113..38cd4197d4 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-6b-chat.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-6b-chat.yaml @@ -3,9 +3,7 @@ label: en_US: 01-ai/Yi-1.5-6B-Chat model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 4096 diff --git a/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-9b-chat.yaml b/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-9b-chat.yaml index 64be8998c5..042eeea81a 100644 --- a/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-9b-chat.yaml +++ b/api/core/model_runtime/model_providers/siliconflow/llm/yi-1.5-9b-chat.yaml @@ -1,11 +1,9 @@ model: 01-ai/Yi-1.5-9B-Chat label: - en_US: 01-ai/Yi-1.5-9B-Chat + en_US: 01-ai/Yi-1.5-9B-Chat-16K model_type: llm features: - - multi-tool-call - agent-thought - - stream-tool-call model_properties: mode: chat context_size: 16384 diff --git a/api/core/model_runtime/model_providers/siliconflow/siliconflow.py b/api/core/model_runtime/model_providers/siliconflow/siliconflow.py index 63f76fa8b5..a53f16c929 100644 --- a/api/core/model_runtime/model_providers/siliconflow/siliconflow.py +++ b/api/core/model_runtime/model_providers/siliconflow/siliconflow.py @@ -19,7 +19,7 @@ class SiliconflowProvider(ModelProvider): model_instance = self.get_model_instance(ModelType.LLM) model_instance.validate_credentials( - model='deepseek-ai/deepseek-v2-chat', + model='deepseek-ai/DeepSeek-V2-Chat', credentials=credentials ) except CredentialsValidateFailedError as ex: diff --git a/api/tests/integration_tests/model_runtime/siliconflow/__init__.py b/api/tests/integration_tests/model_runtime/siliconflow/__init__.py new file mode 100644 index 0000000000..e69de29bb2 diff --git a/api/tests/integration_tests/model_runtime/siliconflow/test_llm.py b/api/tests/integration_tests/model_runtime/siliconflow/test_llm.py new file mode 100644 index 0000000000..befdd82352 --- /dev/null +++ b/api/tests/integration_tests/model_runtime/siliconflow/test_llm.py @@ -0,0 +1,106 @@ +import os +from collections.abc import Generator + +import pytest + +from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta +from core.model_runtime.entities.message_entities import AssistantPromptMessage, SystemPromptMessage, UserPromptMessage +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.siliconflow.llm.llm import SiliconflowLargeLanguageModel + + +def test_validate_credentials(): + model = SiliconflowLargeLanguageModel() + + with pytest.raises(CredentialsValidateFailedError): + model.validate_credentials( + model='deepseek-ai/DeepSeek-V2-Chat', + credentials={ + 'api_key': 'invalid_key' + } + ) + + model.validate_credentials( + model='deepseek-ai/DeepSeek-V2-Chat', + credentials={ + 'api_key': os.environ.get('API_KEY') + } + ) + + +def test_invoke_model(): + model = SiliconflowLargeLanguageModel() + + response = model.invoke( + model='deepseek-ai/DeepSeek-V2-Chat', + credentials={ + 'api_key': os.environ.get('API_KEY') + }, + prompt_messages=[ + UserPromptMessage( + content='Who are you?' + ) + ], + model_parameters={ + 'temperature': 0.5, + 'max_tokens': 10 + }, + stop=['How'], + stream=False, + user="abc-123" + ) + + assert isinstance(response, LLMResult) + assert len(response.message.content) > 0 + + +def test_invoke_stream_model(): + model = SiliconflowLargeLanguageModel() + + response = model.invoke( + model='deepseek-ai/DeepSeek-V2-Chat', + credentials={ + 'api_key': os.environ.get('API_KEY') + }, + prompt_messages=[ + UserPromptMessage( + content='Hello World!' + ) + ], + model_parameters={ + 'temperature': 0.5, + 'max_tokens': 100, + 'seed': 1234 + }, + stream=True, + user="abc-123" + ) + + assert isinstance(response, Generator) + + for chunk in response: + assert isinstance(chunk, LLMResultChunk) + assert isinstance(chunk.delta, LLMResultChunkDelta) + assert isinstance(chunk.delta.message, AssistantPromptMessage) + assert len(chunk.delta.message.content) > 0 if chunk.delta.finish_reason is None else True + + +def test_get_num_tokens(): + model = SiliconflowLargeLanguageModel() + + num_tokens = model.get_num_tokens( + model='deepseek-ai/DeepSeek-V2-Chat', + credentials={ + 'api_key': os.environ.get('API_KEY') + }, + prompt_messages=[ + SystemPromptMessage( + content='You are a helpful AI assistant.', + ), + UserPromptMessage( + content='Hello World!' + ) + ] + ) + + assert num_tokens == 12 diff --git a/api/tests/integration_tests/model_runtime/siliconflow/test_provider.py b/api/tests/integration_tests/model_runtime/siliconflow/test_provider.py new file mode 100644 index 0000000000..7b9211a5db --- /dev/null +++ b/api/tests/integration_tests/model_runtime/siliconflow/test_provider.py @@ -0,0 +1,21 @@ +import os + +import pytest + +from core.model_runtime.errors.validate import CredentialsValidateFailedError +from core.model_runtime.model_providers.siliconflow.siliconflow import SiliconflowProvider + + +def test_validate_provider_credentials(): + provider = SiliconflowProvider() + + with pytest.raises(CredentialsValidateFailedError): + provider.validate_provider_credentials( + credentials={} + ) + + provider.validate_provider_credentials( + credentials={ + 'api_key': os.environ.get('API_KEY') + } + )