Merge pull request #59 from deepset-ai/mistral-changes

Change Mixtral to Mistral

README.md

@@ -19,7 +19,7 @@ For more examples, you may also find our [Blog](https://haystack.deepset.ai/blog
 | Hacker News RAG with Custom Component | <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/hackernews-custom-component-rag.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
 | Use Chroma for RAG and Indexing | <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/chroma-indexing-and-rag-examples.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
 | Using the Jina-embeddings-v2-base-en model in a Haystack RAG pipeline for legal document analsysis| <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/jina-embeddings-v2-legal-analysis-rag.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
-| Multilingual RAG from a podcast with Whisper, Qdrant and Mixtral| <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/multilingual_rag_podcast.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
+| Multilingual RAG from a podcast with Whisper, Qdrant and Mistral| <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/multilingual_rag_podcast.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
 | Improve retrieval by embedding meaningful metadata| <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/improve-retrieval-by-embedding-metadata.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
 | Information extraction via LLMs (Gorilla OpenFunctions)| <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/information-extraction-gorilla.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|
 | Using AstraDB as a data store in your Haystack pipelines| <a href="https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/astradb_haystack_integration.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>|

notebooks/chroma-indexing-and-rag-examples.ipynb

@@ -118,7 +118,7 @@
  "\n",
  "\n",
  "In this example we are using:\n",
- "- The `HuggingFaceTGIGenerator` with the Mistral 8x7B model. (You will need a Hugging Face token to use this model). You can repleace this with any of the other [`Generators`](https://docs.haystack.deepset.ai/v2.0/docs/generators)\n",
+ "- The `HuggingFaceTGIGenerator` with the Mistral-7B-Instruct-v0.1. (You will need a Hugging Face token to use this model). You can repleace this with any of the other [`Generators`](https://docs.haystack.deepset.ai/v2.0/docs/generators)\n",
  "- The `PromptBuilder` which holds the prompt template. You can adjust this to a prompt of your choice\n",
  "- The `ChromaQueryRetriver` which expects a list of queries and retieves the `top_k` most relevant documents from your Chroma collection."
  ]
@@ -161,7 +161,7 @@
  "\"\"\"\n",
  "prompt_builder = PromptBuilder(template=prompt)\n",
  "\n",
- "llm = HuggingFaceTGIGenerator(model=\"mistralai/Mixtral-8x7B-Instruct-v0.1\")\n",
+ "llm = HuggingFaceTGIGenerator(model=\"mistralai/Mistral-7B-Instruct-v0.1\")\n",
  "llm.warm_up()\n",
  "retriever = ChromaQueryTextRetriever(document_store)\n",
  "\n",

notebooks/jina-embeddings-v2-legal-analysis-rag.ipynb

@@ -16,7 +16,7 @@
  "- the [`jina-embeddings-v2-base-en`](https://arxiv.org/abs/2310.19923) model\n",
  "- [Haystack](https://haystack.deepset.ai/), the open source LLM orchestration framework, version [2.0-beta](https://docs.haystack.deepset.ai/v2.0/docs)\n",
  "- [Chroma](https://docs.trychroma.com/getting-started) to store our vector embeddings, via the [Chroma Document Store Haystack integration](https://haystack.deepset.ai/integrations/chroma-documentstore)\n",
- "- the open source [Mixtral 8x7B LLM](https://huggingface.co/docs/transformers/model_doc/mixtral)\n",
+ "- the open source [Mistral 7B Instruct LLM](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1)\n",
  "\n",
  "\n",
  "## Prerequisites:\n",
@@ -191,7 +191,7 @@
  "source": [
  "# Query pipeline\n",
  "\n",
- "Now the real fun begins. Let's create a query pipeline so we can actually start asking questions. We write a prompt allowing us to pass our documents to the Mixtral-8x7B LLM. Then we initiatialize the LLM via the `HuggingFaceTGIGenerator`.\n",
+ "Now the real fun begins. Let's create a query pipeline so we can actually start asking questions. We write a prompt allowing us to pass our documents to the Mistral-7B LLM. Then we initiatialize the LLM via the `HuggingFaceTGIGenerator`.\n",
  "\n",
  "In Haystack 2.0 `retriever`s are tightly coupled to `DocumentStores`. If we pass in the `retriever` we initialized earlier, this pipeline can access those embeddings we generated, and pass them to the LLM."
  ]
@@ -221,7 +221,7 @@
  "\"\"\"\n",
  "\n",
  "text_embedder = JinaTextEmbedder(model=\"jina-embeddings-v2-base-en\")\n",
- "generator = HuggingFaceTGIGenerator(\"mistralai/Mixtral-8x7B-Instruct-v0.1\")\n",
+ "generator = HuggingFaceTGIGenerator(\"mistralai/Mistral-7B-Instruct-v0.1\")\n",
  "generator.warm_up()\n",
  "\n",
  "prompt_builder = PromptBuilder(template=prompt)\n",

notebooks/mixtral-8x7b-for-web-qa.ipynb

@@ -47,7 +47,8 @@
  "source": [
  "## Prompt the Model - Standalone\n",
  "\n",
- "We are using Hugging Face [Text Generation interface](https://github.com/huggingface/text-generation-inference) (TGI) here. This required an API key 👇"
+ "We are using Hugging Face [Text Generation interface](https://github.com/huggingface/text-generation-inference) (TGI) here. This requires an API key 👇 \n",
+ "> Keep in mind that `mistralai/Mixtral-8x7B-Instruct-v0.1` model is not available on the free tier of Hugging Face Inference API."
  ]
  },
  {

notebooks/multilingual_rag_podcast.ipynb

@@ -21,7 +21,7 @@
  "- OpenAI Whisper model for audio transcription\n",
  "- Qdrant vector database\n",
  "- multilingual embedding model: multilingual-e5-large\n",
- "- multilingual LLM: Mixtral-8x7B-Instruct-v0.1"
+ "- multilingual LLM: Mistral-7B-Instruct-v0.1"
  ]
  },
  {
@@ -991,9 +991,9 @@
  }
  ],
  "source": [
- "# load the model (https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1) and try the Generator\n",
+ "# load the model (https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) and try the Generator\n",
  "\n",
- "generator = HuggingFaceTGIGenerator(\"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n",
+ "generator = HuggingFaceTGIGenerator(\"mistralai/Mistral-7B-Instruct-v0.1\",\n",
  " generation_kwargs={\"max_new_tokens\":500})\n",
  "generator.warm_up()\n",
  "\n",