base on Gemma 2B with 10M context length using Infini-attention. # Gemma 2B - 10M Context Gemma 2B with recurrent local attention with context length of up to 10M. Our implementation uses **<32GB** of memory!  **Features:** - 10M sequence length on Gemma 2B. - Runs on less than 32GB of memory. - Native inference optimized for cuda. - Recurrent local attention for O(N) memory. ## Quick Start > **Note:** This is a very early checkpoint of the model. Only 200 steps. We plan on training for a lot more tokens! Install the requirements: ```bash pip install -r requirements.txt ``` Install the model from huggingface - [Huggingface Model](https://huggingface.co/mustafaaljadery/gemma-10M-safetensor). ```bash python main.py ``` Change the `main.py` inference code to the specific prompt you desire. ```python model_path = "./models/gemma-2b-10m" tokenizer = AutoTokenizer.from_pretrained(model_path) model = GemmaForCausalLM.from_pretrained( model_path, torch_dtype=torch.bfloat16 ) prompt_text = "Summarize this harry potter book..." with torch.no_grad(): generated_text = generate( model, tokenizer, prompt_text, max_length=512, temperature=0.8 ) print(generated_text) ``` ## How does this work? The largest bottleneck (in terms of memory) for LLMs is the KV cache. It grows quadratically in vanilla multi-head attention, thus limiting the size of your sequence length. Our approach splits the attention in local attention blocks as outlined by [InfiniAttention](https://arxiv.org/abs/2404.07143). We take those local attention blocks and apply recurrance to the local attention blocks for the final result of 10M context global atention. A lot of the inspiration for our ideas comes from the [Transformer-XL](https://arxiv.org/abs/1901.02860) paper. ## More details For more context about our motivations, implementation details, and the theory behind the work, check out our technical overview on [medium](https://medium.com/@akshgarg_36829/gemma-10m-technical-overview-900adc4fbeeb). ## Credits This was built by: - [Mustafa Aljadery](https://www.maxaljadery.com/) - [Siddharth Sharma](https://stanford.edu/~sidshr/) - [Aksh Garg](https://www.linkedin.com/in/aksh-garg/) ", Assign "at most 3 tags" to the expected json: {"id":"10070","tags":[]} "only from the tags list I provide: [{"id":77,"name":"3d"},{"id":89,"name":"agent"},{"id":17,"name":"ai"},{"id":54,"name":"algorithm"},{"id":24,"name":"api"},{"id":44,"name":"authentication"},{"id":3,"name":"aws"},{"id":27,"name":"backend"},{"id":60,"name":"benchmark"},{"id":72,"name":"best-practices"},{"id":39,"name":"bitcoin"},{"id":37,"name":"blockchain"},{"id":1,"name":"blog"},{"id":45,"name":"bundler"},{"id":58,"name":"cache"},{"id":21,"name":"chat"},{"id":49,"name":"cicd"},{"id":4,"name":"cli"},{"id":64,"name":"cloud-native"},{"id":48,"name":"cms"},{"id":61,"name":"compiler"},{"id":68,"name":"containerization"},{"id":92,"name":"crm"},{"id":34,"name":"data"},{"id":47,"name":"database"},{"id":8,"name":"declarative-gui "},{"id":9,"name":"deploy-tool"},{"id":53,"name":"desktop-app"},{"id":6,"name":"dev-exp-lib"},{"id":59,"name":"dev-tool"},{"id":13,"name":"ecommerce"},{"id":26,"name":"editor"},{"id":66,"name":"emulator"},{"id":62,"name":"filesystem"},{"id":80,"name":"finance"},{"id":15,"name":"firmware"},{"id":73,"name":"for-fun"},{"id":2,"name":"framework"},{"id":11,"name":"frontend"},{"id":22,"name":"game"},{"id":81,"name":"game-engine "},{"id":23,"name":"graphql"},{"id":84,"name":"gui"},{"id":91,"name":"http"},{"id":5,"name":"http-client"},{"id":51,"name":"iac"},{"id":30,"name":"ide"},{"id":78,"name":"iot"},{"id":40,"name":"json"},{"id":83,"name":"julian"},{"id":38,"name":"k8s"},{"id":31,"name":"language"},{"id":10,"name":"learning-resource"},{"id":33,"name":"lib"},{"id":41,"name":"linter"},{"id":28,"name":"lms"},{"id":16,"name":"logging"},{"id":76,"name":"low-code"},{"id":90,"name":"message-queue"},{"id":42,"name":"mobile-app"},{"id":18,"name":"monitoring"},{"id":36,"name":"networking"},{"id":7,"name":"node-version"},{"id":55,"name":"nosql"},{"id":57,"name":"observability"},{"id":46,"name":"orm"},{"id":52,"name":"os"},{"id":14,"name":"parser"},{"id":74,"name":"react"},{"id":82,"name":"real-time"},{"id":56,"name":"robot"},{"id":65,"name":"runtime"},{"id":32,"name":"sdk"},{"id":71,"name":"search"},{"id":63,"name":"secrets"},{"id":25,"name":"security"},{"id":85,"name":"server"},{"id":86,"name":"serverless"},{"id":70,"name":"storage"},{"id":75,"name":"system-design"},{"id":79,"name":"terminal"},{"id":29,"name":"testing"},{"id":12,"name":"ui"},{"id":50,"name":"ux"},{"id":88,"name":"video"},{"id":20,"name":"web-app"},{"id":35,"name":"web-server"},{"id":43,"name":"webassembly"},{"id":69,"name":"workflow"},{"id":87,"name":"yaml"}]" returns me the "expected json"