Disable CUDA: Space runs on CPU basic, no libcuda.so.1

Fix: add tokenizers and transformers packages for tokenizer conversion

Fix subprocess pipe deadlock: use subprocess.run with communicate()

Raise default model size limit to 25GB (9B models work on free tier)

Make model size limit configurable via MAX_MODEL_GB env var

Add model size gate: reject models >14GB on free tier to prevent OOM crashes

Fix: per-file download progress, scaled timeouts for large models

Add legacy mtmd pipeline: LLaVA, MiniCPM-V, GLM-Edge surgery+convert support

Add legacy mtmd pipeline: LLaVA, MiniCPM-V, GLM-Edge surgery+convert support

Add legacy mtmd pipeline: LLaVA, MiniCPM-V, GLM-Edge surgery+convert support

Add legacy mtmd pipeline: LLaVA, MiniCPM-V, GLM-Edge surgery+convert support

Complete BlackRiver dark theme + mmproj fixes + tests

Complete BlackRiver dark theme + mmproj fixes + tests

Fix: body_colour -> body_text_color, remove show_api deprecation

BlackRiver rebrand: fix all bugs, add BlackRiver theme, progress indicators, rebrand to BlackRiver GGUF Tools

BlackRiver rebrand: fix all bugs, add BlackRiver theme, progress indicators, rebrand to BlackRiver GGUF Tools

BlackRiver rebrand: fix all bugs, add BlackRiver theme, progress indicators, rebrand to BlackRiver GGUF Tools

BlackRiver rebrand: fix all bugs, add BlackRiver theme, progress indicators, rebrand to BlackRiver GGUF Tools

BlackRiver rebrand: fix all bugs, add BlackRiver theme, progress indicators, rebrand to BlackRiver GGUF Tools

Upload app.py with huggingface_hub

Update app.py

Update app.py

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload start.sh with huggingface_hub

Upload start.sh with huggingface_hub

Upload 11 files

initial commit