サイト内検索

AI モデル 84755

• yolov10n↗

  jameslahm

• 10Eros-Max-beta6-W4A8-ConvRot↗

  JoaoZaokk

  *The showcase above was rendered entirely at home on two RTX 3090s with the files in this repo: OpenAI's Sora launch "Tokyo walk" prompt, then two made-up characters, picture and sound from the model, about 6 minutes per 5-second 1280×736 clip. All characters are fictional.* Quantizations of TenStrip/10Eros-Max `10Eros_Max_h3_8stepTURBO_beta6_pruned.safetensors` (MiniMax H3, 50 blocks + 2 token-refiner blocks), plus the MiniMax H3 text encoder (Qwen3-VL 32B) in a form that runs natively on Ampere. The diffusion files go in `models/diffusion_models`, the text encoder in `models/text_encoders`. The VAEs are the official ones from Comfy-Org/MiniMax-H3: `minimax_h3_video_vae_fp16` and `minimax_h3_audio_vae_fp32`. Every tensor outside the quantized Linears is the source's, byte for byte.

• Qwen3.8-Flash-Next-MTP-C2T8-refit-GGUF↗

  Jrouth

• realcartoonXL-v6↗

  jzli

• medgemma-4b-it-MLX-4bit↗

  lmstudio-community

• Kimodo-SOMA-SEED-v1.1-GGML↗

  LocalAI-io

• Qwen3.8-27B-MLX-2bit↗

  majentik

• LFM2.5-2.6B-Wolof-Instruct-v3-GGUF↗

  mamelles

• CTFsolver-Q8_0-GGUF↗

  MarkBruzon

• GLM-5.2-colibri-E8-IQ3-with-int8-mtp↗

  mastouri