From 4419d357c4cfd87c2aee7e95895881d2075c1c5f Mon Sep 17 00:00:00 2001 From: Su Yang Date: Tue, 19 Mar 2024 20:54:31 +0800 Subject: [PATCH] chore: update Yi models params (#2895) --- .../yi/llm/yi-34b-chat-0205.yaml | 27 +++++++++++---- .../yi/llm/yi-34b-chat-200k.yaml | 33 ++++++++++++++----- .../model_providers/yi/llm/yi-vl-plus.yaml | 27 +++++++++++---- 3 files changed, 66 insertions(+), 21 deletions(-) diff --git a/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-0205.yaml b/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-0205.yaml index 4d4148aa91..429c646b77 100644 --- a/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-0205.yaml +++ b/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-0205.yaml @@ -9,18 +9,33 @@ model_properties: mode: chat context_size: 4096 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 控制生成结果的多样性和随机性。数值越小，越严谨；数值越大，越发散。 + en_US: Control the diversity and randomness of generated results. The smaller the value, the more rigorous it is; the larger the value, the more divergent it is. - name: max_tokens use_template: max_tokens type: int default: 512 min: 1 - max: 4096 - - name: temperature - use_template: temperature + max: 4000 + 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 type: float - default: 0.7 - min: 0 - max: 2 + default: 0.8 + min: 0.01 + max: 1.00 + help: + zh_Hans: 控制生成结果的随机性。数值越小，随机性越弱；数值越大，随机性越强。一般而言，top_p 和 temperature 两个参数选择一个进行调整即可。 + en_US: Control the randomness of generated results. The smaller the value, the weaker the randomness; the larger the value, the stronger the randomness. Generally speaking, you can adjust one of the two parameters top_p and temperature. pricing: input: '0.0025' output: '0.0025' diff --git a/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-200k.yaml b/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-200k.yaml index 4fbe84e9b7..d0e181d007 100644 --- a/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-200k.yaml +++ b/api/core/model_runtime/model_providers/yi/llm/yi-34b-chat-200k.yaml @@ -9,18 +9,33 @@ model_properties: mode: chat context_size: 200000 parameter_rules: - - name: max_tokens - use_template: max_tokens - type: int - default: 1024 - min: 1 - max: 200000 - name: temperature use_template: temperature type: float - default: 0.7 - min: 0 - max: 2 + default: 0.6 + min: 0.0 + max: 2.0 + help: + zh_Hans: 控制生成结果的多样性和随机性。数值越小，越严谨；数值越大，越发散。 + en_US: Control the diversity and randomness of generated results. The smaller the value, the more rigorous it is; the larger the value, the more divergent it is. + - name: max_tokens + use_template: max_tokens + type: int + default: 4096 + min: 1 + max: 199950 + 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 + type: float + default: 0.9 + min: 0.01 + max: 1.00 + help: + zh_Hans: 控制生成结果的随机性。数值越小，随机性越弱；数值越大，随机性越强。一般而言，top_p 和 temperature 两个参数选择一个进行调整即可。 + en_US: Control the randomness of generated results. The smaller the value, the weaker the randomness; the larger the value, the stronger the randomness. Generally speaking, you can adjust one of the two parameters top_p and temperature. pricing: input: '0.012' output: '0.012' diff --git a/api/core/model_runtime/model_providers/yi/llm/yi-vl-plus.yaml b/api/core/model_runtime/model_providers/yi/llm/yi-vl-plus.yaml index 6195051f16..a6abcc401f 100644 --- a/api/core/model_runtime/model_providers/yi/llm/yi-vl-plus.yaml +++ b/api/core/model_runtime/model_providers/yi/llm/yi-vl-plus.yaml @@ -9,18 +9,33 @@ model_properties: mode: chat context_size: 4096 parameter_rules: + - name: temperature + use_template: temperature + type: float + default: 0.3 + min: 0.0 + max: 2.0 + help: + zh_Hans: 控制生成结果的多样性和随机性。数值越小，越严谨；数值越大，越发散。 + en_US: Control the diversity and randomness of generated results. The smaller the value, the more rigorous it is; the larger the value, the more divergent it is. - name: max_tokens use_template: max_tokens type: int default: 512 min: 1 - max: 4096 - - name: temperature - use_template: temperature + max: 4000 + 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 type: float - default: 0.7 - min: 0 - max: 2 + default: 0.8 + min: 0.01 + max: 1.00 + help: + zh_Hans: 控制生成结果的随机性。数值越小，随机性越弱；数值越大，随机性越强。一般而言，top_p 和 temperature 两个参数选择一个进行调整即可。 + en_US: Control the randomness of generated results. The smaller the value, the weaker the randomness; the larger the value, the stronger the randomness. Generally speaking, you can adjust one of the two parameters top_p and temperature. pricing: input: '0.01' output: '0.03'