diff --git a/gallery/index.yaml b/gallery/index.yaml index 6403b025..578a7e34 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -25,6 +25,25 @@ - cpu - moe - granite +- !!merge <<: *granite3 + name: "moe-girl-800ma-3bt" + icon: https://huggingface.co/allura-org/MoE-Girl-800MA-3BT/resolve/main/moe-girl-800-3.png + url: "github:mudler/LocalAI/gallery/chatml.yaml@master" + urls: + - https://huggingface.co/allura-org/MoE-Girl-800MA-3BT + - https://huggingface.co/mradermacher/MoE-Girl-800MA-3BT-GGUF + description: | + A roleplay-centric finetune of IBM's Granite 3.0 3B-A800M. LoRA finetune trained locally, whereas the others were FFT; while this results in less uptake of training data, it should also mean less degradation in Granite's core abilities, making it potentially easier to use for general-purpose tasks. + Disclaimer + + PLEASE do not expect godliness out of this, it's a model with 800 million active parameters. Expect something more akin to GPT-3 (the original, not GPT-3.5.) (Furthermore, this version is by a less experienced tuner; it's my first finetune that actually has decent-looking graphs, I don't really know what I'm doing yet!) + overrides: + parameters: + model: MoE-Girl-800MA-3BT.Q4_K_M.gguf + files: + - filename: MoE-Girl-800MA-3BT.Q4_K_M.gguf + sha256: 4c3cb57c27aadabd05573a1a01d6c7aee0f21620db919c7704f758d172e0bfa3 + uri: huggingface://mradermacher/MoE-Girl-800MA-3BT-GGUF/MoE-Girl-800MA-3BT.Q4_K_M.gguf - name: "moe-girl-1ba-7bt-i1" icon: https://cdn-uploads.huggingface.co/production/uploads/634262af8d8089ebaefd410e/kTXXSSSqpb21rfyOX7FUa.jpeg # chatml