github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-06-22 14:49:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: update nvidia docs v2 (#21288)

Browse Source 
More doc updates por favor @baskaryan!
This commit is contained in:
Daniel Glogowski 2024-05-06 11:29:02 -07:00 committed by GitHub
parent 6feddfae88
commit 27e73ebe57
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 7 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/docs/integrations/providers/nvidia.mdx
View File

@ -6,7 +6,7 @@
> [NVIDIA AI Foundation Endpoints](https://www.nvidia.com/en-us/ai-data-science/foundation-models/) give users easy access to NVIDIA hosted API endpoints for > [NVIDIA AI Foundation Endpoints](https://www.nvidia.com/en-us/ai-data-science/foundation-models/) give users easy access to NVIDIA hosted API endpoints for
> NVIDIA AI Foundation Models like `Mixtral 8x7B`, `Llama 2`, `Stable Diffusion`, etc. These models, > NVIDIA AI Foundation Models like `Mixtral 8x7B`, `Llama 2`, `Stable Diffusion`, etc. These models,
> hosted on the [NVIDIA NGC catalog](https://catalog.ngc.nvidia.com/ai-foundation-models), are optimized, tested, and hosted on > hosted on the [NVIDIA API catalog](https://build.nvidia.com/), are optimized, tested, and hosted on
> the NVIDIA AI platform, making them fast and easy to evaluate, further customize, > the NVIDIA AI platform, making them fast and easy to evaluate, further customize,
> and seamlessly run at peak performance on any accelerated stack. > and seamlessly run at peak performance on any accelerated stack.
> >

22
docs/docs/integrations/text_embedding/nvidia_ai_endpoints.ipynb
View File

@ -85,9 +85,6 @@
"import getpass\n", "import getpass\n",
"import os\n", "import os\n",
"\n", "\n",
"## API Key can be found by going to NVIDIA NGC -> AI Foundation Models -> (some model) -> Get API Code or similar.\n",
"## 10K free queries to any endpoint (which is a lot actually).\n",
"\n",
"# del os.environ['NVIDIA_API_KEY'] ## delete key and reset\n", "# del os.environ['NVIDIA_API_KEY'] ## delete key and reset\n",
"if os.environ.get(\"NVIDIA_API_KEY\", \"\").startswith(\"nvapi-\"):\n", "if os.environ.get(\"NVIDIA_API_KEY\", \"\").startswith(\"nvapi-\"):\n",
" print(\"Valid NVIDIA_API_KEY already in environment. Delete to reset\")\n", " print(\"Valid NVIDIA_API_KEY already in environment. Delete to reset\")\n",
@ -112,11 +109,7 @@
"source": [ "source": [
"## Initialization\n", "## Initialization\n",
"\n", "\n",
"The main requirement when initializing an embedding model is to provide the model name. An example is `nvolveqa_40k` below.\n", "When initializing an embedding model you can select a model by passing it, e.g. `ai-embed-qa-4` below, or use the default by not passing any arguments."
"\n",
"For `nvovleqa_40k`, you can also specify the `model_type` as `passage` or `query`. When doing retrieval, you will get best results if you embed the source documents with the `passage` type and the user queries with the `query` type.\n",
"\n",
"If not provided, the `embed_query` method will default to the `query` type, and the `embed_documents` mehod will default to the `passage` type."
] ]
}, },
{ {
@ -129,10 +122,7 @@
"source": [ "source": [
"from langchain_nvidia_ai_endpoints import NVIDIAEmbeddings\n", "from langchain_nvidia_ai_endpoints import NVIDIAEmbeddings\n",
"\n", "\n",
"embedder = NVIDIAEmbeddings(model=\"nvolveqa_40k\")\n", "embedder = NVIDIAEmbeddings(model=\"ai-embed-qa-4\")"
"\n",
"# Alternatively, if you want to specify whether it will use the query or passage type\n",
"# embedder = NVIDIAEmbeddings(model=\"nvolveqa_40k\", model_type=\"passage\")"
] ]
}, },
{ {
@ -156,7 +146,7 @@
"id": "pcDu3v4CbmWk" "id": "pcDu3v4CbmWk"
}, },
"source": [ "source": [
"### **Similarity/Speed Test**\n", "### **Similarity**\n",
"\n", "\n",
"The following is a quick test of the methods in terms of usage, format, and speed for the use case of embedding the following data points:\n", "The following is a quick test of the methods in terms of usage, format, and speed for the use case of embedding the following data points:\n",
"\n", "\n",
@ -250,7 +240,7 @@
"s = time.perf_counter()\n", "s = time.perf_counter()\n",
"# To use the \"query\" mode, we have to add it as an instance arg\n", "# To use the \"query\" mode, we have to add it as an instance arg\n",
"q_embeddings = NVIDIAEmbeddings(\n", "q_embeddings = NVIDIAEmbeddings(\n",
" model=\"nvolveqa_40k\", model_type=\"query\"\n", " model=\"ai-embed-qa-4\", model_type=\"query\"\n",
").embed_documents(\n", ").embed_documents(\n",
" [\n", " [\n",
" \"What's the weather like in Komchatka?\",\n", " \"What's the weather like in Komchatka?\",\n",
@ -501,7 +491,7 @@
"source": [ "source": [
"vectorstore = FAISS.from_texts(\n", "vectorstore = FAISS.from_texts(\n",
" [\"harrison worked at kensho\"],\n", " [\"harrison worked at kensho\"],\n",
" embedding=NVIDIAEmbeddings(model=\"nvolveqa_40k\"),\n", " embedding=NVIDIAEmbeddings(model=\"ai-embed-qa-4\"),\n",
")\n", ")\n",
"retriever = vectorstore.as_retriever()\n", "retriever = vectorstore.as_retriever()\n",
"\n", "\n",
@ -515,7 +505,7 @@
" ]\n", " ]\n",
")\n", ")\n",
"\n", "\n",
"model = ChatNVIDIA(model=\"mixtral_8x7b\")\n", "model = ChatNVIDIA(model=\"ai-mixtral-8x7b-instruct\")\n",
"\n", "\n",
"chain = (\n", "chain = (\n",
" {\"context\": retriever, \"question\": RunnablePassthrough()}\n", " {\"context\": retriever, \"question\": RunnablePassthrough()}\n",