--- license: other model-index: - name: Gemmasutra-Mini-2B-v1 results: - task: type: text-generation name: Text Generation dataset: name: IFEval (0-Shot) type: HuggingFaceH4/ifeval args: num_few_shot: 0 metrics: - type: inst_level_strict_acc and prompt_level_strict_acc value: 25.49 name: strict accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=TheDrummer/Gemmasutra-Mini-2B-v1 name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: BBH (3-Shot) type: BBH args: num_few_shot: 3 metrics: - type: acc_norm value: 9.81 name: normalized accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=TheDrummer/Gemmasutra-Mini-2B-v1 name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MATH Lvl 5 (4-Shot) type: hendrycks/competition_math args: num_few_shot: 4 metrics: - type: exact_match value: 3.17 name: exact match source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=TheDrummer/Gemmasutra-Mini-2B-v1 name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: GPQA (0-shot) type: Idavidrein/gpqa args: num_few_shot: 0 metrics: - type: acc_norm value: 2.8 name: acc_norm source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=TheDrummer/Gemmasutra-Mini-2B-v1 name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MuSR (0-shot) type: TAUR-Lab/MuSR args: num_few_shot: 0 metrics: - type: acc_norm value: 1.19 name: acc_norm source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=TheDrummer/Gemmasutra-Mini-2B-v1 name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MMLU-PRO (5-shot) type: TIGER-Lab/MMLU-Pro config: main split: test args: num_few_shot: 5 metrics: - type: acc value: 11.72 name: accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=TheDrummer/Gemmasutra-Mini-2B-v1 name: Open LLM Leaderboard --- # Join our Discord! https://discord.gg/Nbv9pQ88Xb ### Works on [Kobold 1.72](https://github.com/LostRuins/koboldcpp/releases/tag/v1.72) and [Layla (iOS / Android)](https://www.layla-network.ai/) --- ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/HjVYV2h_YTL9P-insb7fz.png) [BeaverAI](https://huggingface.co/BeaverAI) team proudly presents # Gemmasutra Mini 2B v1 🧘 *Gemma's been training for ALL of you.* ![image/webp](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/w0Oi8TReoQNT3ljm5Wf6c.webp) *A tiny RP model packing an unbelievable PUNCH. Finetuned by yours truly.* ## Description Gone are the days when models below 7B were too small to give you a satisfying RP experience. Gemmasutra Mini 2B v1 marks the beginning of a new chapter in our local LLM community. Be it in your browser, your crappy laptop, your mid-tier phone running [Layla](https://www.layla-network.ai/), or your Raspberry Pi... This lil 2B will give you a playthrough worth having. A model for thee, a model for all! *(And yes, it is uncensored and unaligned. Enjoy!)* ## Links - Original: https://huggingface.co/TheDrummer/Gemmasutra-Mini-2B-v1 - GGUF: https://huggingface.co/TheDrummer/Gemmasutra-Mini-2B-v1-GGUF - iMatrix (recommended / better quants): https://huggingface.co/MarsupialAI/Gemmasutra-Mini-2B-v1_iMatrix_GGUF - Mobile (ARM) (Q4_0_X_X): https://huggingface.co/SicariusSicariiStuff/TheDrummer_Gemmasutra-Mini-2B-v1_ARM ## Usage - For the best experience, use the Gemma Instruct template and modify it to support the `system` role (e.g., `system`) - Chat Completion works well too - Don't use it for Math ## Examples A big thanks to @kurgan1138 for most of the logs! ### SFW ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/IstS_eSV_Z52jII1UkZ7d.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/eLU-AlpTd8VCI9H7gOmX2.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/Rnpwj1XE0CflOnOyJtJSK.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/7Y9x9A2HYoyfMkrBpBeO4.png) #### 4K context ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/te_wnC9Tcg2BVeLrtARjr.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/oTiXmVmK7KiI26wgHmS1u.png) ### NSFW NSFW NSFW NSFW NSFW ### NSFW NSFW NSFW NSFW NSFW ### NSFW NSFW NSFW NSFW NSFW ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/J9SwuYQBULHY9Wobljv5V.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/qhCgIEAeXJqA0Msq5mAFA.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/RkgOg_mvjADiXgb5859gm.png) #### Forgiveness Meter ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/_NnPS-kwD7ZpoWUa3wvux.png) #### Group Character ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/TxF2YbcUwJOtcqxWSKybZ.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/CQKkSq6lXj7yKuwcoPdeQ.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/4uWEjHylkhgPOu1UTkg8w.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/q0KFuHhh3ybz51DQze89G.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/Ll8CA5RR7ugTi72P2HBb8.png) # [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_TheDrummer__Gemmasutra-Mini-2B-v1) | Metric |Value| |-------------------|----:| |Avg. | 9.03| |IFEval (0-Shot) |25.49| |BBH (3-Shot) | 9.81| |MATH Lvl 5 (4-Shot)| 3.17| |GPQA (0-shot) | 2.80| |MuSR (0-shot) | 1.19| |MMLU-PRO (5-shot) |11.72|