From 668ec2fadc9fa124e73b77cc1577e962bfaa826e Mon Sep 17 00:00:00 2001 From: Ettore Di Giacinto Date: Wed, 13 Nov 2024 09:45:02 +0100 Subject: [PATCH] models(gallery): add eva-qwen2.5-72b-v0.1-i1 (#4136) Signed-off-by: Ettore Di Giacinto --- gallery/index.yaml | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/gallery/index.yaml b/gallery/index.yaml index cd88830b..bd5d31b9 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -1289,6 +1289,25 @@ - filename: calme-3.1-qwenloi-3b.Q5_K_M.gguf sha256: 8962a8d1704979039063b5c69fafdb38b545c26143419ec4c574f37f2d6dd7b2 uri: huggingface://MaziyarPanahi/calme-3.1-qwenloi-3b-GGUF/calme-3.1-qwenloi-3b.Q5_K_M.gguf +- !!merge <<: *qwen25 + name: "eva-qwen2.5-72b-v0.1-i1" + urls: + - https://huggingface.co/EVA-UNIT-01/EVA-Qwen2.5-72B-v0.1 + - https://huggingface.co/mradermacher/EVA-Qwen2.5-72B-v0.1-i1-GGUF + description: | + A RP/storywriting specialist model, full-parameter finetune of Qwen2.5-72B on mixture of synthetic and natural data. + It uses Celeste 70B 0.1 data mixture, greatly expanding it to improve versatility, creativity and "flavor" of the resulting model. + + Dedicated to Nev. + + Version notes for 0.1: Reprocessed dataset (via Cahvay for 32B 0.2, used here as well), readjusted training config for 8xH100 SXM. Significant improvements in instruction following, long context understanding and overall coherence over v0.0. + overrides: + parameters: + model: EVA-Qwen2.5-72B-v0.1.i1-Q4_K_M.gguf + files: + - filename: EVA-Qwen2.5-72B-v0.1.i1-Q4_K_M.gguf + sha256: b05dbc02eeb286c41122b103ac31431fc8dcbd80b8979422541a05cda53df61b + uri: huggingface://mradermacher/EVA-Qwen2.5-72B-v0.1-i1-GGUF/EVA-Qwen2.5-72B-v0.1.i1-Q4_K_M.gguf - &archfunct license: apache-2.0 tags: