JonathanColetti/Qwen3.8-27B-Uncensored-GGUF
Update README.md
Update README.md
Update README.md
Fix remaining vision projector reference
Rename F16 vision projector for automatic discovery
Add tested ComfyUI setup and update vision projector name
Remove redundant BF16 vision projector
Update README.md
list draft-Q4_0 and vision-bf16 in Files and Whats here
add baseline bf16 perplexity against the unmodified base model
rename vision projector and draft head to match repo naming scheme
add mtp-Qwen3.8-27B-Uncensored-Q4_0.gguf
add mmproj-Qwen3.8-27B-Uncensored-BF16.gguf
Update README.md
Update README.md
Update README.md
Card: fix filenames that do not exist (mtp-*, mmproj-*), document f16 build step
Card: add IQ2_M speculative decoding measurements
gitattributes: LFS globs for *.gguf and *.dat, drop stale per-file rules
Card: add IQ2_M, perplexity table, imatrix docs, self-build recipe
Add Qwen3.8-27B-Uncensored-noMTP-IQ2_M.gguf (imatrix, MTP blk.64 pinned q8_0, token_embd q4_K)
Add Qwen3.8-27B-Uncensored-IQ2_M.gguf (imatrix, MTP blk.64 pinned q8_0, token_embd q4_K)
Remove Qwen3.8-27B-Uncensored-noMTP-IQ2_XXS.gguf
Remove Qwen3.8-27B-Uncensored-IQ2_XXS.gguf
Add Qwen3.8-27B-Uncensored-noMTP-IQ2_XXS.gguf (imatrix, MTP blk.64 pinned q8_0, token_embd q4_K)
Add Qwen3.8-27B-Uncensored-IQ2_XXS.gguf (imatrix, MTP blk.64 pinned q8_0, token_embd q4_K)
Add imatrix (wikitext-2 raw test, 200x512, computed from f16)
Add measured capability benchmarks vs base model
Upload Qwen3.8-27B-Uncensored-vision-f16.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-noMTP-Q8_0.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-noMTP-Q6_K.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-noMTP-Q5_K_M.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-noMTP-Q4_K_M.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-noMTP-IQ4_XS.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-draft-Q8_0.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-Q8_0.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-Q6_K.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-Q5_K_M.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-Q4_K_M.gguf with huggingface_hub
Upload Qwen3.8-27B-Uncensored-IQ4_XS.gguf with huggingface_hub
Upload README.md with huggingface_hub
Archive measured refusal/KL front
Archive Heretic optuna study (full Pareto search)
Delete _uploadtest.gguf with huggingface_hub
Upload _uploadtest.gguf with huggingface_hub
Upload README.md with huggingface_hub
initial commit