From b191d828efad649daa966ffe0191d966080fbe40 Mon Sep 17 00:00:00 2001 From: Narendranath Gogineni <narenrockstar1@gmail.com> Date: Mon, 18 Nov 2024 17:27:20 +0530 Subject: [PATCH 1/3] add support for huggingface on vertex --- src/providers/google-vertex-ai/api.ts | 4 ++++ src/providers/google-vertex-ai/index.ts | 12 ++++++++++++ 2 files changed, 16 insertions(+) diff --git a/src/providers/google-vertex-ai/api.ts b/src/providers/google-vertex-ai/api.ts index 3076e5268..f79777b53 100644 --- a/src/providers/google-vertex-ai/api.ts +++ b/src/providers/google-vertex-ai/api.ts @@ -86,6 +86,10 @@ export const GoogleApiConfig: ProviderAPIConfig = { return `${projectRoute}/endpoints/openapi/chat/completions`; } + case 'huggingface': { + return `${projectRoute}/endpoints/openapi/chat/completions`; + } + default: return `${projectRoute}`; } diff --git a/src/providers/google-vertex-ai/index.ts b/src/providers/google-vertex-ai/index.ts index 95e5e01c1..042969cd1 100644 --- a/src/providers/google-vertex-ai/index.ts +++ b/src/providers/google-vertex-ai/index.ts @@ -17,6 +17,8 @@ import { GoogleImageGenConfig, GoogleImageGenResponseTransform, } from './imageGenerate'; +import { chatCompleteParams, responseTransformers } from '../open-ai-base'; +import { GOOGLE_VERTEX_AI } from '../../globals'; const VertexConfig: ProviderConfigs = { api: VertexApiConfig, @@ -57,6 +59,16 @@ const VertexConfig: ProviderConfigs = { 'stream-chatComplete': VertexLlamaChatCompleteStreamChunkTransform, }, }; + case 'huggingface': + return { + chatComplete: chatCompleteParams([], { + model: 'meta-llama-3-8b-instruct', + }), + api: GoogleApiConfig, + responseTransforms: responseTransformers(GOOGLE_VERTEX_AI, { + chatComplete: true, + }), + }; } }, }; From cd3d04fad038e02015ae33c35f5ff430b7ac982a Mon Sep 17 00:00:00 2001 From: Narendranath Gogineni <narenrockstar1@gmail.com> Date: Fri, 22 Nov 2024 17:48:15 +0530 Subject: [PATCH 2/3] changes for vertex dedicated deployments --- src/providers/google-vertex-ai/api.ts | 4 ++-- src/providers/google-vertex-ai/utils.ts | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/providers/google-vertex-ai/api.ts b/src/providers/google-vertex-ai/api.ts index f79777b53..486854acd 100644 --- a/src/providers/google-vertex-ai/api.ts +++ b/src/providers/google-vertex-ai/api.ts @@ -86,8 +86,8 @@ export const GoogleApiConfig: ProviderAPIConfig = { return `${projectRoute}/endpoints/openapi/chat/completions`; } - case 'huggingface': { - return `${projectRoute}/endpoints/openapi/chat/completions`; + case 'endpoints': { + return `${projectRoute}/endpoints/${model}/chat/completions`; } default: diff --git a/src/providers/google-vertex-ai/utils.ts b/src/providers/google-vertex-ai/utils.ts index adc897228..ccb9673f2 100644 --- a/src/providers/google-vertex-ai/utils.ts +++ b/src/providers/google-vertex-ai/utils.ts @@ -130,7 +130,7 @@ export const getModelAndProvider = (modelString: string) => { const modelStringParts = modelString.split('.'); if ( modelStringParts.length > 1 && - ['google', 'anthropic', 'meta'].includes(modelStringParts[0]) + ['google', 'anthropic', 'meta', 'endpoints'].includes(modelStringParts[0]) ) { provider = modelStringParts[0]; model = modelStringParts.slice(1).join('.'); From 0445b97e4fcfa8093f73986c48e256cf7f2b11be Mon Sep 17 00:00:00 2001 From: Narendranath Gogineni <narenrockstar1@gmail.com> Date: Fri, 22 Nov 2024 18:38:32 +0530 Subject: [PATCH 3/3] minor fix --- src/providers/google-vertex-ai/index.ts | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/providers/google-vertex-ai/index.ts b/src/providers/google-vertex-ai/index.ts index 042969cd1..a087fb926 100644 --- a/src/providers/google-vertex-ai/index.ts +++ b/src/providers/google-vertex-ai/index.ts @@ -59,7 +59,7 @@ const VertexConfig: ProviderConfigs = { 'stream-chatComplete': VertexLlamaChatCompleteStreamChunkTransform, }, }; - case 'huggingface': + case 'endpoints': return { chatComplete: chatCompleteParams([], { model: 'meta-llama-3-8b-instruct',