diff --git a/docs.json b/docs.json index ed4d996..80ec89a 100644 --- a/docs.json +++ b/docs.json @@ -36,7 +36,7 @@ "logo": { "light": "/logo/light.svg", "dark": "/logo/dark.svg", - "href": "/" + "href": "https://liquid.ai" }, "navbar": { "links": [ @@ -52,20 +52,6 @@ } }, "navigation": { - "global": { - "anchors": [ - { - "anchor": "About Us", - "icon": "building", - "href": "https://www.liquid.ai/company/about" - }, - { - "anchor": "Blog", - "icon": "pencil", - "href": "https://www.liquid.ai/company/blog" - } - ] - }, "tabs": [ { "tab": "Documentation", @@ -202,7 +188,7 @@ ] }, { - "tab": "Guides", + "tab": "Examples", "groups": [ { "group": "Get Started", diff --git a/docs/fine-tuning/leap-finetune.mdx b/docs/fine-tuning/leap-finetune.mdx index f7b7c11..d25add6 100644 --- a/docs/fine-tuning/leap-finetune.mdx +++ b/docs/fine-tuning/leap-finetune.mdx @@ -20,10 +20,10 @@ LEAP Finetune will provide: While LEAP Finetune is in development, you can fine-tune models using: - + Hugging Face's training library with LoRA/QLoRA support - + Memory-efficient fine-tuning with 2x faster training diff --git a/docs/frameworks/outlines.mdx b/docs/frameworks/outlines.mdx index 750b6ff..1984a1e 100644 --- a/docs/frameworks/outlines.mdx +++ b/docs/frameworks/outlines.mdx @@ -19,7 +19,7 @@ pip install outlines transformers torch ## Setup[​](#setup "Direct link to Setup") -Outlines provides a simple interface for constrained generation. The examples below use Transformers, but Outlines works with all major inference frameworks including [vLLM](/lfm/inference/vllm), [llama.cpp](/lfm/inference/llama-cpp), [MLX](/lfm/inference/mlx), [Ollama](/lfm/inference/ollama), and more. See the [Outlines documentation](https://dottxt-ai.github.io/outlines/latest/) for framework-specific examples. +Outlines provides a simple interface for constrained generation. The examples below use Transformers, but Outlines works with all major inference frameworks including [vLLM](/docs/inference/vllm), [llama.cpp](/docs/inference/llama-cpp), [MLX](/docs/inference/mlx), [Ollama](/docs/inference/ollama), and more. See the [Outlines documentation](https://dottxt-ai.github.io/outlines/latest/) for framework-specific examples. Start by wrapping your model: @@ -263,5 +263,3 @@ For a detailed example of using Outlines with LFM2-350M for smart home control, * [Outlines GitHub](https://github.com/dottxt-ai/outlines) * [Outlines Documentation](https://dottxt-ai.github.io/outlines/) * [LFM2 × .txt Collaboration Blog Post](https://www.liquid.ai/blog/liquid-txt-collaboration) - -[Edit this page](https://github.com/Liquid4All/docs/tree/main/lfm/frameworks/outlines.md) diff --git a/docs/help/contributing.mdx b/docs/help/contributing.mdx index 31e0bc1..0f84180 100644 --- a/docs/help/contributing.mdx +++ b/docs/help/contributing.mdx @@ -102,8 +102,8 @@ Use Mintlify components appropriately: ### Links -- Use relative links for internal pages: `/lfm/inference/transformers` -- Use descriptive link text: "See the [inference guide](/lfm/inference/transformers)" not "Click [here](/lfm/inference/transformers)" +- Use relative links for internal pages: `/docs/inference/transformers` +- Use descriptive link text: "See the [inference guide](/docs/inference/transformers)" not "Click [here](/docs/inference/transformers)" ## What to Contribute diff --git a/docs/help/faqs.mdx b/docs/help/faqs.mdx index 2a0b8d2..4610e6f 100644 --- a/docs/help/faqs.mdx +++ b/docs/help/faqs.mdx @@ -15,11 +15,11 @@ All LFM models support a 32k token text context length for extended conversation LFM models are compatible with: -- [Transformers](/lfm/inference/transformers) - For research and development -- [llama.cpp](/lfm/inference/llama-cpp) - For efficient CPU inference -- [vLLM](/lfm/inference/vllm) - For high-throughput production serving -- [MLX](/lfm/inference/mlx) - For Apple Silicon optimization -- [Ollama](/lfm/inference/ollama) - For easy local deployment +- [Transformers](/docs/inference/transformers) - For research and development +- [llama.cpp](/docs/inference/llama-cpp) - For efficient CPU inference +- [vLLM](/docs/inference/vllm) - For high-throughput production serving +- [MLX](/docs/inference/mlx) - For Apple Silicon optimization +- [Ollama](/docs/inference/ollama) - For easy local deployment - [LEAP](/leap/index) - For edge and mobile deployment @@ -39,7 +39,7 @@ LFM2.5 models are updated versions with improved training that deliver higher pe -[Liquid Nanos](/lfm/models/liquid-nanos) are task-specific models fine-tuned for specialized use cases like: +[Liquid Nanos](/docs/models/liquid-nanos) are task-specific models fine-tuned for specialized use cases like: - Information extraction (LFM2-Extract) - Translation (LFM2-350M-ENJP-MT) - RAG question answering (LFM2-1.2B-RAG) @@ -69,7 +69,7 @@ For most use cases, Q4_K_M or Q5_K_M provide good quality with significant size ## Fine-tuning -Yes! Most LFM models support fine-tuning with [TRL](/lfm/fine-tuning/trl) and [Unsloth](/lfm/fine-tuning/unsloth). Check the [Complete Model Library](/lfm/models/complete-library) for trainability information. +Yes! Most LFM models support fine-tuning with [TRL](/docs/fine-tuning/trl) and [Unsloth](/docs/fine-tuning/unsloth). Check the [Model Library](/docs/models/complete-library) for trainability information. @@ -82,4 +82,4 @@ Yes! Most LFM models support fine-tuning with [TRL](/lfm/fine-tuning/trl) and [U - Join our [Discord community](https://discord.gg/DFU3WQeaYD) for real-time help - Check the [Cookbook](https://github.com/Liquid4All/cookbook) for examples -- See [Troubleshooting](/lfm/help/troubleshooting) for common issues +- See [Troubleshooting](/docs/help/troubleshooting) for common issues diff --git a/docs/index.mdx b/docs/index.mdx index 9fcb65b..456d6a7 100644 --- a/docs/index.mdx +++ b/docs/index.mdx @@ -3,6 +3,6 @@ title: "LFM Documentation" description: "Redirect to LFM Getting Started" --- - + -Redirecting to [Getting Started](/lfm/getting-started/intro)... +Redirecting to [Getting Started](/docs/getting-started/welcome)... diff --git a/docs/inference/llama-cpp.mdx b/docs/inference/llama-cpp.mdx index 2d95ab5..f6a8294 100644 --- a/docs/inference/llama-cpp.mdx +++ b/docs/inference/llama-cpp.mdx @@ -114,67 +114,25 @@ hf download LiquidAI/LFM2.5-1.2B-Instruct-GGUF lfm2.5-1.2b-instruct-q4_k_m.gguf ## Basic Usage -llama.cpp offers three main interfaces for running inference: `llama-cpp-python` (Python bindings), `llama-server` (OpenAI-compatible server), and `llama-cli` (interactive CLI). +llama.cpp offers two main interfaces for running inference: `llama-server` (OpenAI-compatible server) and `llama-cli` (interactive CLI). - - For Python applications, use the `llama-cpp-python` package. - - **Installation:** - ```bash - pip install llama-cpp-python - ``` - - For GPU support: - ```bash - CMAKE_ARGS="-DLLAMA_CUDA=on" pip install llama-cpp-python - ``` - - **Model Setup:** - ```python - from llama_cpp import Llama - - # Load model - llm = Llama( - model_path="lfm2.5-1.2b-instruct-q4_k_m.gguf", - n_ctx=4096, - n_threads=8 - ) - - # Generate text - output = llm( - "What is artificial intelligence?", - max_tokens=512, - temperature=0.7, - top_p=0.9 - ) - print(output["choices"][0]["text"]) - ``` - - **Chat Completions:** - ```python - response = llm.create_chat_completion( - messages=[ - {"role": "system", "content": "You are a helpful assistant."}, - {"role": "user", "content": "Explain quantum computing."} - ], - temperature=0.7, - max_tokens=512 - ) - print(response["choices"][0]["message"]["content"]) - ``` - - llama-server provides an OpenAI-compatible API for serving models locally. **Starting the Server:** ```bash + llama-server -hf LiquidAI/LFM2.5-1.2B-Instruct-GGUF -c 4096 --port 8080 + ``` + + The `-hf` flag downloads the model directly from Hugging Face. Alternatively, use a local model file: + ```bash llama-server -m lfm2.5-1.2b-instruct-q4_k_m.gguf -c 4096 --port 8080 ``` Key parameters: - * `-m`: Path to GGUF model file + * `-hf`: Hugging Face model ID (downloads automatically) + * `-m`: Path to local GGUF model file * `-c`: Context length (default: 4096) * `--port`: Server port (default: 8080) * `-ngl 99`: Offload layers to GPU (if available) @@ -216,12 +174,18 @@ llama.cpp offers three main interfaces for running inference: `llama-cpp-python` llama-cli provides an interactive terminal interface for chatting with models. + ```bash + llama-cli -hf LiquidAI/LFM2.5-1.2B-Instruct-GGUF -c 4096 --color -i + ``` + + The `-hf` flag downloads the model directly from Hugging Face. Alternatively, use a local model file: ```bash llama-cli -m lfm2.5-1.2b-instruct-q4_k_m.gguf -c 4096 --color -i ``` Key parameters: - * `-m`: Path to GGUF model file + * `-hf`: Hugging Face model ID (downloads automatically) + * `-m`: Path to local GGUF model file * `-c`: Context length * `--color`: Colored output * `-i`: Interactive mode @@ -242,43 +206,6 @@ Control text generation behavior using parameters in the OpenAI-compatible API o * **`repetition_penalty`** / **`--repeat-penalty`** (`float`, default 1.1): Penalty for repeating tokens (>1.0 = discourage repetition). Typical range: 1.0-1.5 * **`stop`** (`str` or `list[str]`): Strings that terminate generation when encountered - - ```python - from llama_cpp import Llama - - llm = Llama( - model_path="lfm2.5-1.2b-instruct-q4_k_m.gguf", - n_ctx=4096, - n_threads=8 - ) - - # Text generation with sampling parameters - output = llm( - "What is machine learning?", - max_tokens=512, - temperature=0.7, - top_p=0.9, - top_k=40, - repeat_penalty=1.1, - stop=["<|im_end|>", "<|endoftext|>"] - ) - print(output["choices"][0]["text"]) - - # Chat completion with sampling parameters - response = llm.create_chat_completion( - messages=[ - {"role": "user", "content": "Explain quantum computing."} - ], - temperature=0.7, - top_p=0.9, - top_k=40, - max_tokens=512, - repeat_penalty=1.1 - ) - print(response["choices"][0]["message"]["content"]) - ``` - - ```python from openai import OpenAI @@ -407,38 +334,6 @@ hf download LiquidAI/LFM2-VL-1.6B-GGUF mmproj-LFM2-VL-1.6B-Q8_0.gguf --local-dir ``` - - ```python - from llama_cpp import Llama - from llama_cpp.llama_chat_format import Llava15ChatHandler - - # Initialize with vision support - # Note: Use the correct chat handler for your model architecture - chat_handler = Llava15ChatHandler(clip_model_path="mmproj-model-f16.gguf") - - llm = Llama( - model_path="lfm2.5-vl-1.6b-q4_k_m.gguf", - chat_handler=chat_handler, - n_ctx=4096 - ) - - # Generate with image - response = llm.create_chat_completion( - messages=[ - { - "role": "user", - "content": [ - {"type": "image_url", "image_url": {"url": "file:///path/to/image.jpg"}}, - {"type": "text", "text": "Describe this image."} - ] - } - ], - max_tokens=256 - ) - print(response["choices"][0]["message"]["content"]) - ``` - - For a complete working example with step-by-step instructions, see the [llama.cpp Vision Model Colab notebook](https://colab.research.google.com/drive/1q2PjE6O_AahakRlkTNJGYL32MsdUcj7b?usp=sharing). diff --git a/docs/inference/transformers.mdx b/docs/inference/transformers.mdx index 3129191..a8509e4 100644 --- a/docs/inference/transformers.mdx +++ b/docs/inference/transformers.mdx @@ -14,9 +14,14 @@ Transformers provides the most flexibility for model development and is ideal fo Install the required dependencies: ```bash -pip install transformers>=4.57.1 torch>=2.6 +pip install "transformers>=5.0.0" torch ``` +> **Note:** Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: +> ```bash +> pip install git+https://github.com/huggingface/transformers.git@0c9a72e4576fe4c84077f066e585129c97bfd4e6 torch +> ``` + GPU is recommended for faster inference. ## Basic Usage diff --git a/docs/inference/vllm.mdx b/docs/inference/vllm.mdx index 21da2b0..c327514 100644 --- a/docs/inference/vllm.mdx +++ b/docs/inference/vllm.mdx @@ -185,9 +185,16 @@ To use LFM Vision Models with vLLM, install the precompiled wheel along with the VLLM_PRECOMPILED_WHEEL_COMMIT=72506c98349d6bcd32b4e33eec7b5513453c1502 VLLM_USE_PRECOMPILED=1 pip install git+https://github.com/vllm-project/vllm.git ``` +```bash +pip install "transformers>=5.0.0" pillow +``` + + +Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow ``` + This installs vLLM with the necessary changes for LFM Vision Model support. Once these changes are merged upstream, you'll be able to use the standard vLLM installation. diff --git a/docs/key-concepts/models.mdx b/docs/key-concepts/models.mdx index 3609c36..e8ca0e2 100644 --- a/docs/key-concepts/models.mdx +++ b/docs/key-concepts/models.mdx @@ -5,7 +5,7 @@ description: "The LFM model collection includes general-purpose language models, * These models are built on the backbone of a new hybrid architecture that's designed for incredibly fast training and inference. Learn more in our [blog post](https://www.liquid.ai/blog/liquid-foundation-models-v2-our-second-series-of-generative-ai-models). * All models support a **32k token text context length** for extended conversations and document processing. -* Our models are compatible with various open-source deployment libraries including [Transformers](/lfm/inference/transformers), [llama.cpp](/lfm/inference/llama-cpp), [vLLM](/lfm/inference/vllm), [MLX](/lfm/inference/mlx), [Ollama](/lfm/inference/ollama), and our own edge deployment platform [LEAP](/lfm/frameworks/leap). +* Our models are compatible with various open-source deployment libraries including [Transformers](/docs/inference/transformers), [llama.cpp](/docs/inference/llama-cpp), [vLLM](/docs/inference/vllm), [MLX](/docs/inference/mlx), [Ollama](/docs/inference/ollama), and our own edge deployment platform [LEAP](/leap/index). | Model | HF | GGUF | MLX | ONNX | Trainable? | @@ -119,25 +119,25 @@ description: "The LFM model collection includes general-purpose language models, | Model | Description | | ----------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| [`LiquidAI/LFM2-1.2B-Extract`](https://huggingface.co/LiquidAI/LFM2-1.2B-Extract) | Extract important information from a wide variety of unstructured documents into structured outputs like JSON. [See prompting guidelines](/lfm/key-concepts/text-generation-and-prompting#lfm2-extract) | -| [`LiquidAI/LFM2-350M-Extract`](https://huggingface.co/LiquidAI/LFM2-350M-Extract) | Smaller version of the extraction model. [See prompting guidelines](/lfm/key-concepts/text-generation-and-prompting#lfm2-extract) | -| [`LiquidAI/LFM2-350M-ENJP-MT`](https://huggingface.co/LiquidAI/LFM2-350M-ENJP-MT) | Near real-time bi-directional Japanese/English translation of short-to-medium inputs. > [See prompting guidelines](/lfm/key-concepts/text-generation-and-prompting#lfm2-350m-enjp-mt) | -| [`LiquidAI/LFM2-1.2B-RAG`](https://huggingface.co/LiquidAI/LFM2-1.2B-RAG) | Answer questions based on provided contextual documents, for use in RAG systems. > [See prompting guidelines](/lfm/key-concepts/text-generation-and-prompting#lfm2-rag) | +| [`LiquidAI/LFM2-1.2B-Extract`](https://huggingface.co/LiquidAI/LFM2-1.2B-Extract) | Extract important information from a wide variety of unstructured documents into structured outputs like JSON. [See prompting guidelines](/docs/key-concepts/text-generation-and-prompting#lfm2-extract) | +| [`LiquidAI/LFM2-350M-Extract`](https://huggingface.co/LiquidAI/LFM2-350M-Extract) | Smaller version of the extraction model. [See prompting guidelines](/docs/key-concepts/text-generation-and-prompting#lfm2-extract) | +| [`LiquidAI/LFM2-350M-ENJP-MT`](https://huggingface.co/LiquidAI/LFM2-350M-ENJP-MT) | Near real-time bi-directional Japanese/English translation of short-to-medium inputs. > [See prompting guidelines](/docs/key-concepts/text-generation-and-prompting#lfm2-350m-enjp-mt) | +| [`LiquidAI/LFM2-1.2B-RAG`](https://huggingface.co/LiquidAI/LFM2-1.2B-RAG) | Answer questions based on provided contextual documents, for use in RAG systems. > [See prompting guidelines](/docs/key-concepts/text-generation-and-prompting#lfm2-rag) | | [`LiquidAI/LFM2-350M-Math`](https://huggingface.co/LiquidAI/LFM2-350M-Math) | Tiny reasoning model designed for tackling tricky math problems. | -| [`LiquidAI/LFM2-350M-PII-Extract-JP`](https://huggingface.co/LiquidAI/LFM2-350M-PII-Extract-JP) | Extract personally identifiable information (PII) from Japanese text and output it in JSON format. [See prompting guidelines](/lfm/key-concepts/text-generation-and-prompting#lfm2-350m-pii-extract-jp) | +| [`LiquidAI/LFM2-350M-PII-Extract-JP`](https://huggingface.co/LiquidAI/LFM2-350M-PII-Extract-JP) | Extract personally identifiable information (PII) from Japanese text and output it in JSON format. [See prompting guidelines](/docs/key-concepts/text-generation-and-prompting#lfm2-350m-pii-extract-jp) | | [`LiquidAI/LFM2-ColBERT-350M`](https://huggingface.co/LiquidAI/LFM2-ColBERT-350M) | Embed documents and queries for fast retrieval and reranking across many languages. | -| [`LiquidAI/LFM2-2.6B-Transcript`](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript) | Designed for private, on-device meeting summarization. [See prompting guidelines](/lfm/key-concepts/text-generation-and-prompting#lfm2-2.6b-transcript) | +| [`LiquidAI/LFM2-2.6B-Transcript`](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript) | Designed for private, on-device meeting summarization. [See prompting guidelines](/docs/key-concepts/text-generation-and-prompting#lfm2-2.6b-transcript) | | [`LiquidAI/LFM2-1.2B-Tool`](https://huggingface.co/LiquidAI/LFM2-1.2B-Tool) | Deprecated Model optimized for concise and precise tool calling. See updated [`LFM2.5-1.2B-Instruct`](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct) instead. | ## GGUF Models[​](#gguf-models "Direct link to GGUF Models") -GGUF quantized versions are available for all LFM2 models for efficient inference with [llama.cpp](/lfm/inference/llama-cpp), [LM Studio](/lfm/inference/lm-studio), and [Ollama](/lfm/inference/ollama). These models offer reduced memory usage and faster CPU inference. +GGUF quantized versions are available for all LFM2 models for efficient inference with [llama.cpp](/docs/inference/llama-cpp), [LM Studio](/docs/inference/lm-studio), and [Ollama](/docs/inference/ollama). These models offer reduced memory usage and faster CPU inference. To access our official GGUF models, append `-GGUF` to any model repository name (e.g., `LiquidAI/LFM2-1.2B-GGUF`). All models are available in multiple quantization levels (`Q4_0`, `Q4_K_M`, `Q5_K_M`, `Q6_K`, `Q8_0`, `F16`). ## MLX Models[​](#mlx-models "Direct link to MLX Models") -MLX quantized versions are available for many of the LFM2 model library for efficient inference on Apple Silicon with [MLX](/lfm/inference/mlx). These models leverage unified memory architecture for optimal performance on M-series chips. +MLX quantized versions are available for many of the LFM2 model library for efficient inference on Apple Silicon with [MLX](/docs/inference/mlx). These models leverage unified memory architecture for optimal performance on M-series chips. Browse all MLX-compatible models at [mlx-community LFM2 models](https://huggingface.co/mlx-community/collections?search=LFM). All models are available in multiple quantization levels (`4-bit`, `5-bit`, `6-bit`, `8-bit`, `bf16`). @@ -146,5 +146,3 @@ Browse all MLX-compatible models at [mlx-community LFM2 models](https://huggingf ONNX versions are available for many LFM2 models for cross-platform deployment and inference with ONNX Runtime. These models enable deployment across diverse hardware including CPUs, GPUs, and specialized accelerators. To access our official ONNX models, append `-ONNX` to any model repository name (e.g., `LiquidAI/LFM2.5-1.2B-Instruct-ONNX`). - -[Edit this page](https://github.com/Liquid4All/docs/tree/main/lfm/key-concepts/models.md) diff --git a/docs/key-concepts/text-generation-and-prompting.mdx b/docs/key-concepts/text-generation-and-prompting.mdx index 365540c..2cba4ef 100644 --- a/docs/key-concepts/text-generation-and-prompting.mdx +++ b/docs/key-concepts/text-generation-and-prompting.mdx @@ -90,8 +90,6 @@ Parameter names and syntax vary by platform. See [Transformers](/docs/inference/ * `min_p=0.15` * `repetition_penalty=1.05` -Note that Liquid Nanos have [special requirements](#liquid-nanos) with different parameters. - ## Vision Models LFM2-VL models use a **variable resolution encoder** to control the quality/speed tradeoff by adjusting how images are tokenized. @@ -133,319 +131,6 @@ min_image_tokens = 32 * `max_image_tokens=256` * `do_image_splitting=True` -## Liquid Nanos - -### LFM2-Extract - -Structured information extraction models. Use `temperature=0` (greedy decoding). - -**System Prompt Format:** - -``` -Identify and extract information matching the following schema. -Return data as a JSON object. Missing data should be omitted. - -Schema: -- field_name: "Description of what to extract" -- nested_object: - - nested_field: "Description" -``` - -If no system prompt is provided, defaults to JSON. Specify format (JSON, XML, or YAML) and schema for better accuracy. - - - **System Prompt:** - - ``` - Identify and extract information matching the following schema. - Return data as a JSON object. Missing data should be omitted. - - Schema: - - project_name: "The name of the martial arts venture" - - executive_summary: "A brief overview of the business" - - financial_projections: - - startup_costs: "Initial investment required" - - projected_revenue_year_1: "Expected revenue in the first year" - - projected_revenue_year_3: "Expected revenue in the third year" - - return_on_investment: "Timeframe for recouping investment" - - facility_details: - - location: - - address: "Full address of the facility" - - latitude: "Latitude coordinate" - - longitude: "Longitude coordinate" - - distance_to_park_boundary: "Distance to park boundary" - - noise_level_assessment: "Noise level rating" - ``` - - **User Prompt:** - - ``` - We have secured a lease for a facility located at 345 Ocean View Drive, - Seward, AK 99664. The facility offers 3000 sq ft of dedicated training - space. Its location is approximately 2 miles from the boundary of the - Kenai Fjords National Park. A noise level assessment has been conducted, - resulting in a rating of Moderate. The coordinates of the facility are - 60.1728° N latitude and 149.3422° W longitude. - ``` - - **Assistant Response:** - - ```json - { - "project_name": "Arctic Spirit Martial Arts", - "executive_summary": "This business plan outlines the establishment of a premier martial arts training center in Seward, Alaska.", - "financial_projections": { - "startup_costs": "$75000", - "projected_revenue_year_1": "$120000", - "projected_revenue_year_3": "$250000", - "return_on_investment": "3 years" - }, - "facility_details": { - "location": { - "address": "345 Ocean View Drive, Seward, AK 99664", - "latitude": "60.1728° N", - "longitude": "149.3422° W", - "distance_to_park_boundary": "2 miles", - "noise_level_assessment": "Moderate" - } - } - } - ``` - - -⚠️ **The model is intended for single turn conversations.** - -### LFM2-RAG - -Optimized for Retrieval-Augmented Generation. Use `temperature=0` (greedy decoding). Provide relevant documents in the system prompt. - -**System Prompt Format:** - -``` -The following documents may provide you additional information to answer questions: - - -[Document content here] - - - -[Document content here] - -``` - - - **System Prompt:** - - ``` - The following documents may provide you additional information to answer questions: - - - The centre, which was created in 1906, has been instrumental in advancing - agriculture research. The library at the Agriculture Canada research centre - in Lethbridge serves 48 scientists and 85 technicians, along with many - visiting staff and students. - - ``` - - **User Prompt:** - - ``` - How many individuals were reported to be served by the library at the - Agriculture Canada research centre in Lethbridge? - ``` - - **Assistant Response:** - - ``` - The library at the Agriculture Canada research centre in Lethbridge was - reported to serve 48 scientists and 85 technicians, along with many - visiting staff and students. - ``` - - -### LFM2-Tool - -Optimized for efficient and precise tool calling. Use `temperature=0` (greedy decoding). See the [Tool Use](/docs/key-concepts/tool-use) guide for details. - -### LFM2-350M-ENJP-MT - -Specialized English-Japanese translation model. **Requires** one of these system prompts: - -* `"Translate to Japanese."` - English → Japanese -* `"Translate to English."` - Japanese → English - -Use `.apply_chat_template()` with the system prompt to specify direction. - - - ```python - from transformers import AutoTokenizer - - tokenizer = AutoTokenizer.from_pretrained("LiquidAI/LFM2-350M-ENJP-MT") - - chat = [ - {"role": "system", "content": "Translate to Japanese."}, - {"role": "user", "content": "What is C. elegans?"} - ] - - prompt = tokenizer.apply_chat_template(chat, tokenize=False) - print(prompt) - ``` - - **Output:** - - ``` - <|startoftext|><|im_start|>system - Translate to Japanese.<|im_end|> - <|im_start|>user - What is C. elegans?<|im_end|> - <|im_start|>assistant - C. elegansとは何ですか？<|im_end|> - ``` - - -⚠️ **The model is intended for single turn conversations.** - -### LFM2-350M-PII-Extract-JP - -Extracts PII from Japanese text as JSON. Use `temperature=0` (greedy decoding). Output can mask sensitive information on-device. - -**System Prompt Format:** - -``` -Extract
, , , , -``` - -Extract specific entities by listing only what you need (e.g., `Extract `). List categories in alphabetical order for optimal performance. - -**Output Format:** JSON with lists per category. Empty lists for missing entities. Outputs entities exactly as they appear (including notation variations) for exact-match masking. - - - **System Prompt:** - - ``` - Extract
, , , , - ``` - - **User Prompt:** - - ``` - こんにちは、ラミンさんに B200 GPU を 10000 台 至急請求してください。 - 連絡先は celegans@liquid.ai (電話番号010-000-0000) で、これは C. elegans - 線虫に着想を得たニューラルネットワークアーキテクチャを 今すぐ構築するために不可欠です。 - ``` - - **Assistant Response:** - - ```json - { - "address": [], - "company_name": [], - "email_address": ["celegans@liquid.ai"], - "human_name": ["ラミン"], - "phone_number": ["010-000-0000"] - } - ``` - - -⚠️ **The model is intended for single turn conversations.** - -### LFM2-2.6B-Transcript - -Designed for private, on-device meeting summarization. Use `temperature=0.3` for optimal results. - -**Generation Parameters:** - -* `temperature=0.3` (strongly recommended) - -**System Prompt Format:** - -``` -You are an expert meeting analyst. Analyze the transcript carefully -and provide clear, accurate information based on the content. -``` - -**Input Format:** - -The model expects meeting transcripts in a specific format: - -``` - - -Title (example: Claims Processing training module) -Date (example: July 2, 2021) -Time (example: 1:00 PM) -Duration (example: 45 minutes) -Participants (example: Julie Franco (Training Facilitator), Amanda Newman (Subject Matter Expert)) - ----------- - -**Speaker 1**: Message 1 (example: **Julie Franco**: Good morning, everyone. Thanks for joining me today.) -**Speaker 2**: Message 2 (example: **Amanda Newman**: Good morning, Julie. Happy to be here.) -etc. -``` - -Replace `` with one of the following summary types, or combine multiple prompts: - -| Summary type | User prompt | -| ----------------- | ------------------------------------------------------------------------------------------------------------------------------- | -| Executive summary | Provide a brief executive summary (2-3 sentences) of the key outcomes and decisions from this transcript. | -| Detailed summary | Provide a detailed summary of the transcript, covering all major topics, discussions, and outcomes in paragraph form. | -| Action items | List the specific action items that were assigned during this meeting. Include who is responsible for each item when mentioned. | -| Key decisions | List the key decisions that were made during this meeting. Focus on concrete decisions and outcomes. | -| Participants | List the participants mentioned in this transcript. Include their roles or titles when available. | -| Topics discussed | List the main topics and subjects that were discussed in this meeting. | - - - **Example inputs and outputs:** - - | Title | Input meeting | Model output | - | --------------------------- | ----------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- | - | Budget planning | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/meeting1.txt) | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/output1.txt) | - | Design review | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/meeting2.txt) | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/output2.txt) | - | Coffee chat / social hour | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/meeting3.txt) | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/output3.txt) | - | Procurement / vendor review | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/meeting4.txt) | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/output4.txt) | - | Task force meeting | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/meeting5.txt) | [Link](https://huggingface.co/LiquidAI/LFM2-2.6B-Transcript/resolve/main/examples/output5.txt) | - - **System Prompt:** - - ``` - You are an expert meeting analyst. Analyze the transcript carefully - and provide clear, accurate information based on the content. - ``` - - **User Prompt:** - - ``` - Provide a detailed summary of the transcript, covering all major topics, - discussions, and outcomes in paragraph form. - - Title: Budget Planning Meeting - Date: March 15, 2024 - Time: 2:00 PM - Duration: 60 minutes - Participants: Sarah Chen (Finance Director), Mike Johnson (Operations Manager), Lisa Park (HR Manager) - - ---------- - - **Sarah Chen**: Good afternoon everyone. Let's start by reviewing our Q1 budget performance. - **Mike Johnson**: Thanks Sarah. Operations came in 5% under budget this quarter. - **Lisa Park**: HR expenses were slightly over due to the new hiring initiative, but we're on track for Q2. - **Sarah Chen**: That's good to hear. For Q2, we need to allocate additional funds for the expansion project. - **Mike Johnson**: I can provide a detailed breakdown by next week. - **Lisa Park**: I'll coordinate with Mike on staffing needs for the expansion. - ``` - - **Assistant Response:** - - ``` - The Budget Planning Meeting on March 15, 2024, reviewed Q1 financial performance - and planned for Q2. Operations performed well, coming in 5% under budget, while - HR expenses were slightly over due to new hiring. The team discussed allocating - additional funds for an expansion project in Q2, with Mike Johnson committing to - provide a detailed breakdown by next week and Lisa Park coordinating staffing needs. - ``` - - -**Try it yourself:** See the [meeting summarization cookbook example](https://github.com/Liquid4All/cookbook/tree/main/examples/meeting-summarization) for a complete implementation. - -⚠️ **The model is intended for single turn conversations with a specific format.** + +**Liquid Nanos** (task-specific models like LFM2-Extract, LFM2-RAG, LFM2-Tool, etc.) may have special prompting requirements and different generation parameters. For the best usage guidelines, refer to the individual model cards on the [Liquid Nanos](/docs/models/liquid-nanos) page. + diff --git a/docs/models/complete-library.mdx b/docs/models/complete-library.mdx index d1f8383..64a2abd 100644 --- a/docs/models/complete-library.mdx +++ b/docs/models/complete-library.mdx @@ -1,5 +1,5 @@ --- -title: "Complete Model Library" +title: "Model Library" description: "Liquid Foundation Models (LFMs) are a new class of multimodal architectures built for fast inference and on-device deployment. Browse all available models and formats here." --- diff --git a/docs/models/lfm2-8b-a1b.mdx b/docs/models/lfm2-8b-a1b.mdx index f41baae..1a99bc8 100644 --- a/docs/models/lfm2-8b-a1b.mdx +++ b/docs/models/lfm2-8b-a1b.mdx @@ -45,8 +45,15 @@ LFM2-8B-A1B is Liquid AI's Mixture-of-Experts model, combining 8B total paramete **Install:** ```bash + pip install "transformers>=5.0.0" torch + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: + ```bash pip install git+https://github.com/huggingface/transformers.git@0c9a72e4576fe4c84077f066e585129c97bfd4e6 torch ``` + **Download & Run:** ```python diff --git a/docs/models/lfm2-vl-1.6b.mdx b/docs/models/lfm2-vl-1.6b.mdx index db66865..b331a7d 100644 --- a/docs/models/lfm2-vl-1.6b.mdx +++ b/docs/models/lfm2-vl-1.6b.mdx @@ -31,8 +31,15 @@ LFM2-VL-1.6B was the original 1.6B vision-language model. It has been superseded **Install:** ```bash + pip install "transformers>=5.0.0" pillow torch + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: + ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow torch ``` + **Download & Run:** ```python @@ -86,9 +93,16 @@ LFM2-VL-1.6B was the original 1.6B vision-language model. It has been superseded pip install git+https://github.com/vllm-project/vllm.git ``` + ```bash + pip install "transformers>=5.0.0" pillow + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow ``` + **Run:** ```python diff --git a/docs/models/lfm2-vl-3b.mdx b/docs/models/lfm2-vl-3b.mdx index 7f5c71c..009981f 100644 --- a/docs/models/lfm2-vl-3b.mdx +++ b/docs/models/lfm2-vl-3b.mdx @@ -45,8 +45,15 @@ LFM2-VL-3B is Liquid AI's highest-capacity multimodal model, delivering enhanced **Install:** ```bash + pip install "transformers>=5.0.0" pillow torch + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: + ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow torch ``` + **Download & Run:** ```python @@ -100,9 +107,16 @@ LFM2-VL-3B is Liquid AI's highest-capacity multimodal model, delivering enhanced pip install git+https://github.com/vllm-project/vllm.git ``` + ```bash + pip install "transformers>=5.0.0" pillow + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow ``` + **Run:** ```python diff --git a/docs/models/lfm2-vl-450m.mdx b/docs/models/lfm2-vl-450m.mdx index 4393517..ddaa038 100644 --- a/docs/models/lfm2-vl-450m.mdx +++ b/docs/models/lfm2-vl-450m.mdx @@ -45,8 +45,15 @@ LFM2-VL-450M is Liquid AI's smallest vision-language model, designed for edge de **Install:** ```bash + pip install "transformers>=5.0.0" pillow torch + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: + ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow torch ``` + **Download & Run:** ```python @@ -100,9 +107,16 @@ LFM2-VL-450M is Liquid AI's smallest vision-language model, designed for edge de pip install git+https://github.com/vllm-project/vllm.git ``` + ```bash + pip install "transformers>=5.0.0" pillow + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow ``` + **Run:** ```python diff --git a/docs/models/lfm25-vl-1.6b.mdx b/docs/models/lfm25-vl-1.6b.mdx index db79510..61cbccf 100644 --- a/docs/models/lfm25-vl-1.6b.mdx +++ b/docs/models/lfm25-vl-1.6b.mdx @@ -46,8 +46,15 @@ LFM2.5-VL-1.6B is Liquid AI's flagship vision-language model, delivering excepti **Install:** ```bash + pip install "transformers>=5.0.0" pillow torch + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: + ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow torch ``` + **Download & Run:** ```python @@ -101,9 +108,16 @@ LFM2.5-VL-1.6B is Liquid AI's flagship vision-language model, delivering excepti pip install git+https://github.com/vllm-project/vllm.git ``` + ```bash + pip install "transformers>=5.0.0" pillow + ``` + + + Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: ```bash pip install git+https://github.com/huggingface/transformers.git@3c2517727ce28a30f5044e01663ee204deb1cdbe pillow ``` + **Run:** ```python diff --git a/examples/index.mdx b/examples/index.mdx index 192499f..b6c0e61 100644 --- a/examples/index.mdx +++ b/examples/index.mdx @@ -1,5 +1,5 @@ --- -title: "Guides Library" +title: "Examples Library" --- diff --git a/leap/edge-sdk/android/cloud-ai-comparison.mdx b/leap/edge-sdk/android/cloud-ai-comparison.mdx index 87a039f..7f6ca41 100644 --- a/leap/edge-sdk/android/cloud-ai-comparison.mdx +++ b/leap/edge-sdk/android/cloud-ai-comparison.mdx @@ -162,4 +162,4 @@ lifecycleScope.launch { ## Next steps -For more information, please refer to the [quick start guide](android-quick-start-guide) and [API reference](android-api-spec). +For more information, please refer to the [quick start guide](./android-quick-start-guide). diff --git a/leap/edge-sdk/android/constrained-generation.mdx b/leap/edge-sdk/android/constrained-generation.mdx index 5dab90e..f01220e 100644 --- a/leap/edge-sdk/android/constrained-generation.mdx +++ b/leap/edge-sdk/android/constrained-generation.mdx @@ -4,7 +4,7 @@ description: "Generate structured JSON output with compile-time validation using sidebar_position: 3 --- -Setting the `jsonSchemaConstraint` field in [`GenerationOptions`](./android-api-spec#generationoptions) will enable constrained generation. While it is possible to +Setting the `jsonSchemaConstraint` field in [`GenerationOptions`](./conversation-generation#generationoptions) will enable constrained generation. While it is possible to directly set the constraint with raw JSON Schema strings, we recommend to create the constraints with the `Generatable` annotation. ## `Generatable` annotation diff --git a/leap/edge-sdk/android/conversation-generation.mdx b/leap/edge-sdk/android/conversation-generation.mdx index 540a402..7b80ef8 100644 --- a/leap/edge-sdk/android/conversation-generation.mdx +++ b/leap/edge-sdk/android/conversation-generation.mdx @@ -74,7 +74,7 @@ Register a function for the model to invoke. See [function calling](./function-c Export the whole conversation history into a `JSONArray`. Each element can be interpreted as a `ChatCompletionRequestMessage` instance in OpenAI API schema. -See also: [Gson Support](../utilities#gson-support). +See also: [Gson Support](./utilities#gson-support). ### Cancellation of the generation @@ -129,12 +129,12 @@ Fields - `topP`: Nucleus sampling parameter.In nucleus sampling, the model only considers the results of the tokens with `topP` probability mass. - `minP`: Minimal possibility for a token to be considered in generation. - `repetitionPenalty`: Repetition penalty parameter. A positive value will decrease the model's likelihood to repeat the same line verbatim. -- `jsonSchemaConstraint`: Enable constrained generation with a [JSON Schema](https://json-schema.org/). See [constrained generation](../advanced-features#constrained-generation) for more details. +- `jsonSchemaConstraint`: Enable constrained generation with a [JSON Schema](https://json-schema.org/). See [constrained generation](./constrained-generation) for more details. - `functionCallParser`: Define the parser for function calling requests from the model. See [function calling](./function-calling) guide for more details. Methods -- `setResponseFormatType`: Enable constrained generation with a `Generatable` data class. See [constrained generation](../advanced-features#constrained-generation) for more details. +- `setResponseFormatType`: Enable constrained generation with a `Generatable` data class. See [constrained generation](./constrained-generation) for more details. Kotlin builder function `GenerationOptions.build` is also available. For example, diff --git a/leap/edge-sdk/android/function-calling.mdx b/leap/edge-sdk/android/function-calling.mdx index 0afffff..0bfdc81 100644 --- a/leap/edge-sdk/android/function-calling.mdx +++ b/leap/edge-sdk/android/function-calling.mdx @@ -10,7 +10,7 @@ Not all models support function calling. Please check the model card before usin ## Register functions to conversations -To enable function calling, function definitions should be registered to the [`Conversation`](./android-api-spec#conversation) instance before content generation. +To enable function calling, function definitions should be registered to the [`Conversation`](./conversation-generation#conversation) instance before content generation. `Conversation.registerFunction` takes a `LeapFunction` instance as the input, which describes the name, parameters and ability of the function. ```kotlin diff --git a/leap/edge-sdk/android/messages-content.mdx b/leap/edge-sdk/android/messages-content.mdx index 7335527..94b5347 100644 --- a/leap/edge-sdk/android/messages-content.mdx +++ b/leap/edge-sdk/android/messages-content.mdx @@ -31,7 +31,7 @@ ChatMessage.fromJSONObject(obj: JSONObject): ChatMessage Return a `JSONObject` that represents the chat message. The returned object is compatible with `ChatCompletionRequestMessage` from OpenAI API. It contains 2 fields: `role` and `content` . -See also: [Gson Support](../utilities#gson-support). +See also: [Gson Support](./utilities#gson-support). ### `fromJSONObject` @@ -42,7 +42,7 @@ Construct a `ChatMessage` instance from a `JSONObject`. Not all JSON object vari message. -See also: [Gson Support](../utilities#gson-support). +See also: [Gson Support](./utilities#gson-support). ### `ChatMessage.Role` diff --git a/leap/edge-sdk/android/model-loading.mdx b/leap/edge-sdk/android/model-loading.mdx index 0691404..09c54dc 100644 --- a/leap/edge-sdk/android/model-loading.mdx +++ b/leap/edge-sdk/android/model-loading.mdx @@ -30,7 +30,7 @@ Download a model from the LEAP Model Library and load it into memory. If the mod **Returns** -`ModelRunner`: A [`ModelRunner`](../conversation-generation#modelrunner) instance that can be used to interact with the loaded model. +`ModelRunner`: A [`ModelRunner`](./conversation-generation#modelrunner) instance that can be used to interact with the loaded model.
### `downloadModel` @@ -126,7 +126,7 @@ It is rarely necessary to instantiate a `Manifest` class directly. It is created ### `loadModel` - This function can be called from UI thread. The app should hold the `ModelRunner` object returned by this function until there is no need to interact with the model anymore. See [`ModelRunner`](../conversation-generation#modelrunner) for more details. + This function can be called from UI thread. The app should hold the `ModelRunner` object returned by this function until there is no need to interact with the model anymore. See [`ModelRunner`](./conversation-generation#modelrunner) for more details. **Arguments** diff --git a/leap/edge-sdk/android/utilities.mdx b/leap/edge-sdk/android/utilities.mdx index 09903ad..8448780 100644 --- a/leap/edge-sdk/android/utilities.mdx +++ b/leap/edge-sdk/android/utilities.mdx @@ -26,8 +26,8 @@ dependencies { The following types are supported: -- [`ChatMessage`](../messages-content#chatmessage) -- [`ChatMessageContent`](../messages-content#chatmessagecontent) +- [`ChatMessage`](./messages-content#chatmessage) +- [`ChatMessageContent`](./messages-content#chatmessagecontent) ### Create Gson Object @@ -42,7 +42,7 @@ val gson = GsonBuilder().registerLeapAdapters().create() ### Serializing and Deserializing Conversation History -With a [`Conversation`](../conversation-generation#conversation) object, simply call `Gson.toJson` to convert the chat message history into a JSON string. The returned JSON will be an array. +With a [`Conversation`](./conversation-generation#conversation) object, simply call `Gson.toJson` to convert the chat message history into a JSON string. The returned JSON will be an array. ```kotlin val json = gson.toJson(conversation.history) @@ -59,7 +59,7 @@ val chatHistory: List = gson.fromJson(json, LeapGson.messageListTyp ## Model Downloader (deprecated) -This module is deprecated and will be removed in the near future. To download models using the Edge SDK, see [`LeapDownloader`](../model-loading#leapdownloader). +This module is deprecated and will be removed in the near future. To download models using the Edge SDK, see [`LeapDownloader`](./model-loading#leapdownloader). LeapSDK Android Model Downloader module is a helper for downloading models from Leap Model Library. diff --git a/leap/edge-sdk/ios/cloud-ai-comparison.mdx b/leap/edge-sdk/ios/cloud-ai-comparison.mdx index 6beacc9..e4419ac 100644 --- a/leap/edge-sdk/ios/cloud-ai-comparison.mdx +++ b/leap/edge-sdk/ios/cloud-ai-comparison.mdx @@ -178,4 +178,4 @@ func sendMessage(_ text: String) { ## Next steps -For more information, please refer to the [quick start guide](./ios-quick-start-guide) and [API reference](ios-api-spec). +For more information, please refer to the [quick start guide](./ios-quick-start-guide). diff --git a/leap/edge-sdk/ios/function-calling.mdx b/leap/edge-sdk/ios/function-calling.mdx index ceb86a6..401e153 100644 --- a/leap/edge-sdk/ios/function-calling.mdx +++ b/leap/edge-sdk/ios/function-calling.mdx @@ -17,7 +17,7 @@ your messages and tool responses. ## Register functions to conversations -To enable function calling, function definitions should be registered to the [`Conversation`](./ios-api-spec#conversation) instance before content generation. +To enable function calling, function definitions should be registered to the [`Conversation`](./conversation-generation#conversation) instance before content generation. `Conversation.registerFunction` takes a `LeapFunction` instance as the input, which describes the name, parameters and ability of the function. ```swift diff --git a/leap/edge-sdk/ios/model-loading.mdx b/leap/edge-sdk/ios/model-loading.mdx index cc36861..4c56242 100644 --- a/leap/edge-sdk/ios/model-loading.mdx +++ b/leap/edge-sdk/ios/model-loading.mdx @@ -33,7 +33,7 @@ public struct Leap { **Returns** -`ModelRunner`: A [`ModelRunner`](../conversation-generation#modelrunner) instance that can be used to interact with the loaded model. +`ModelRunner`: A [`ModelRunner`](./conversation-generation#modelrunner) instance that can be used to interact with the loaded model.
diff --git a/quickstarts/LFM2-8B-A1B__transformers.md b/quickstarts/LFM2-8B-A1B__transformers.md index 2061940..b891bd5 100644 --- a/quickstarts/LFM2-8B-A1B__transformers.md +++ b/quickstarts/LFM2-8B-A1B__transformers.md @@ -11,9 +11,14 @@ Perfect for research, prototyping, and quick experimentation in Jupyter notebook ## Install Python dependencies ```shell -pip install git+https://github.com/huggingface/transformers.git@0c9a72e4576fe4c84077f066e585129c97bfd4e6 bitsandbytes +pip install "transformers>=5.0.0" bitsandbytes ``` +> **Note:** Transformers v5 is newly released. If you encounter issues, fall back to the pinned git source: +> ```shell +> pip install git+https://github.com/huggingface/transformers.git@0c9a72e4576fe4c84077f066e585129c97bfd4e6 bitsandbytes +> ``` + ## Run inference ```python diff --git a/style.css b/style.css index 712f910..4814ff5 100644 --- a/style.css +++ b/style.css @@ -62,6 +62,12 @@ a[href*="discord.gg"]:hover img { background-color: #864bc4 !important; } +/* Hide external link arrow on Discord card */ +a[href*="discord"] svg[class*="arrow"], +a[href*="discord.gg"] svg[class*="arrow"] { + display: none !important; +} + /* Light mode LEFT sidebar - selected item styling */ /* Target links with bg-primary-light class (Mintlify's active state) */ :root:not(.dark) a[class*="bg-primary-light"], diff --git a/styles.js b/styles.js index 302e5e1..48727bc 100644 --- a/styles.js +++ b/styles.js @@ -84,6 +84,12 @@ background-color: #864bc4 !important; } + /* Hide external link arrow on Discord card */ + a[href*="discord"] svg[class*="arrow"], + a[href*="discord.gg"] svg[class*="arrow"] { + display: none !important; + } + /* Light mode LEFT sidebar - selected item styling */ /* Target links with bg-primary-light class (Mintlify's active state) */ :root:not(.dark) a[class*="bg-primary-light"],