From f8f3c35f99485f65b94ca285c73d26b2310f3974 Mon Sep 17 00:00:00 2001 From: Gianluca Boiano Date: Mon, 20 Jan 2025 13:28:25 +0100 Subject: [PATCH] chore(model gallery): update icons and add LLAVA ones Signed-off-by: Gianluca Boiano --- gallery/index.yaml | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-) diff --git a/gallery/index.yaml b/gallery/index.yaml index 679ab002..1a9c9880 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -2181,7 +2181,6 @@ sha256: 42cf7a96784dc8f25c61c2404620c3e6548a024caa8dff6e435d7c86400d7ab8 uri: huggingface://mradermacher/Qwen2.5-7B-nerd-uncensored-v1.7-GGUF/Qwen2.5-7B-nerd-uncensored-v1.7.Q4_K_M.gguf - !!merge <<: *qwen25 - icon: https://i.imgur.com/OxX2Usi.png name: "evathene-v1.0" urls: - https://huggingface.co/sophosympatheia/Evathene-v1.0 @@ -2540,7 +2539,6 @@ sha256: 91907f29746625a62885793475956220b81d8a5a34b53686a1acd1d03fd403ea uri: huggingface://bartowski/72B-Qwen2.5-Kunou-v1-GGUF/72B-Qwen2.5-Kunou-v1-Q4_K_M.gguf - !!merge <<: *qwen25 - icon: https://i.imgur.com/OxX2Usi.png name: "evathene-v1.3" urls: - https://huggingface.co/sophosympatheia/Evathene-v1.3 @@ -4485,7 +4483,6 @@ sha256: 27b10c3ca4507e8bf7d305d60e5313b54ef5fffdb43a03f36223d19d906e39f3 uri: huggingface://mradermacher/L3.1-70Blivion-v0.1-rc1-70B-i1-GGUF/L3.1-70Blivion-v0.1-rc1-70B.i1-Q4_K_M.gguf - !!merge <<: *llama31 - icon: https://i.imgur.com/sdN0Aqg.jpeg name: "llama-3.1-hawkish-8b" urls: - https://huggingface.co/mukaj/Llama-3.1-Hawkish-8B @@ -5225,7 +5222,7 @@ - &deepseek ## Deepseek url: "github:mudler/LocalAI/gallery/deepseek.yaml@master" name: "deepseek-coder-v2-lite-instruct" - icon: "https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" + icon: "https://avatars.githubusercontent.com/u/148330874" license: deepseek description: | DeepSeek-Coder-V2, an open-source Mixture-of-Experts (MoE) code language model that achieves performance comparable to GPT4-Turbo in code-specific tasks. Specifically, DeepSeek-Coder-V2 is further pre-trained from DeepSeek-Coder-V2-Base with 6 trillion tokens sourced from a high-quality and multi-source corpus. Through this continued pre-training, DeepSeek-Coder-V2 substantially enhances the coding and mathematical reasoning capabilities of DeepSeek-Coder-V2-Base, while maintaining comparable performance in general language tasks. Compared to DeepSeek-Coder, DeepSeek-Coder-V2 demonstrates significant advancements in various aspects of code-related tasks, as well as reasoning and general capabilities. Additionally, DeepSeek-Coder-V2 expands its support for programming languages from 86 to 338, while extending the context length from 16K to 128K. @@ -6155,7 +6152,6 @@ - !!merge <<: *mistral03 name: "mn-12b-mag-mell-r1-iq-arm-imatrix" url: "github:mudler/LocalAI/gallery/chatml.yaml@master" - icon: "https://i.imgur.com/wjyAaTO.png" urls: - https://huggingface.co/inflatebot/MN-12B-Mag-Mell-R1 - https://huggingface.co/Lewdiculous/MN-12B-Mag-Mell-R1-GGUF-IQ-ARM-Imatrix @@ -7265,7 +7261,6 @@ name: "l3-8b-stheno-v3.1" urls: - https://huggingface.co/Sao10K/L3-8B-Stheno-v3.1 - icon: https://w.forfun.com/fetch/cb/cba2205390e517bea1ea60ca0b491af4.jpeg description: | - A model made for 1-on-1 Roleplay ideally, but one that is able to handle scenarios, RPGs and storywriting fine. - Uncensored during actual roleplay scenarios. # I do not care for zero-shot prompting like what some people do. It is uncensored enough in actual usecases. @@ -8059,7 +8054,6 @@ urls: - https://huggingface.co/bartowski/New-Dawn-Llama-3-70B-32K-v1.0-GGUF - https://huggingface.co/sophosympatheia/New-Dawn-Llama-3-70B-32K-v1.0 - icon: https://imgur.com/tKzncGo.png description: | This model is a multi-level SLERP merge of several Llama 3 70B variants. See the merge recipe below for details. I extended the context window for this model out to 32K by snagging some layers from abacusai/Smaug-Llama-3-70B-Instruct-32K using a technique similar to what I used for Midnight Miqu, which was further honed by jukofyork. This model is uncensored. You are responsible for whatever you do with it. @@ -8411,7 +8405,8 @@ - filename: dolphin-2.9.2-Phi-3-Medium-abliterated-Q4_K_M.gguf sha256: 566331c2efe87725310aacb709ca15088a0063fa0ddc14a345bf20d69982156b uri: huggingface://bartowski/dolphin-2.9.2-Phi-3-Medium-abliterated-GGUF/dolphin-2.9.2-Phi-3-Medium-abliterated-Q4_K_M.gguf -- url: "github:mudler/LocalAI/gallery/chatml.yaml@master" +- !!merge <<: *llama3 + url: "github:mudler/LocalAI/gallery/chatml.yaml@master" name: "llama-3-8b-instruct-dpo-v0.3-32k" license: llama3 urls: @@ -8882,6 +8877,8 @@ sha256: 4cc1cb3660d87ff56432ebeb7884ad35d67c48c7b9f6b2856f305e39c38eed8f uri: huggingface://moondream/moondream2-gguf/moondream2-mmproj-f16.gguf - &llava ### START LLaVa + name: "llava-1.6-vicuna" + icon: https://github.com/lobehub/lobe-icons/raw/master/packages/static-png/dark/llava-color.png url: "github:mudler/LocalAI/gallery/llava.yaml@master" license: apache-2.0 description: | @@ -8895,7 +8892,6 @@ - gpu - llama2 - cpu - name: "llava-1.6-vicuna" overrides: mmproj: mmproj-vicuna7b-f16.gguf parameters: @@ -9363,7 +9359,6 @@ June 18, 2024 Update, After extensive testing of the intermediate checkpoints, significant progress has been made. The model is slowly — I mean, really slowly — unlearning its alignment. By significantly lowering the learning rate, I was able to visibly observe deep behavioral changes, this process is taking longer than anticipated, but it's going to be worth it. Estimated time to completion: 4 more days.. I'm pleased to report that in several tests, the model not only maintained its intelligence but actually showed a slight improvement, especially in terms of common sense. An intermediate checkpoint of this model was used to create invisietch/EtherealRainbow-v0.3-rc7, with promising results. Currently, it seems like I'm on the right track. I hope this model will serve as a solid foundation for further merges, whether for role-playing (RP) or for uncensoring. This approach also allows us to save on actual fine-tuning, thereby reducing our carbon footprint. The merge process takes just a few minutes of CPU time, instead of days of GPU work. June 20, 2024 Update, Unaligning was partially successful, and the results are decent, but I am not fully satisfied. I decided to bite the bullet, and do a full finetune, god have mercy on my GPUs. I am also releasing the intermediate checkpoint of this model. - icon: https://i.imgur.com/Kpk1PgZ.png overrides: parameters: model: LLAMA-3_8B_Unaligned_Alpha-Q4_K_M.gguf @@ -9389,7 +9384,6 @@ uri: huggingface://bartowski/L3-8B-Lunaris-v1-GGUF/L3-8B-Lunaris-v1-Q4_K_M.gguf - !!merge <<: *llama3 name: "llama-3_8b_unaligned_alpha_rp_soup-i1" - icon: https://i.imgur.com/pXcjpoV.png urls: - https://huggingface.co/SicariusSicariiStuff/LLAMA-3_8B_Unaligned_Alpha_RP_Soup - https://huggingface.co/mradermacher/LLAMA-3_8B_Unaligned_Alpha_RP_Soup-i1-GGUF @@ -9787,7 +9781,6 @@ sha256: 9c90f3a65332a03a6cbb563eee19c7586d9544f646ff9f33f7f1904b3d415ae2 uri: huggingface://nold/HelpingAI-9B-GGUF/HelpingAI-9B_Q4_K_M.gguf - url: "github:mudler/LocalAI/gallery/chatml-hercules.yaml@master" - icon: "https://tse3.mm.bing.net/th/id/OIG1.vnrl3xpEcypR3McLW63q?pid=ImgGn" urls: - https://huggingface.co/Locutusque/Llama-3-Hercules-5.0-8B - https://huggingface.co/bartowski/Llama-3-Hercules-5.0-8B-GGUF