Spaces:
Runtime error
Runtime error
Commit History
-
Move xformers disable to global scope for better compatibility
Patryk Ptasiński +1: ClaudeThe previous fix didn't work because xformers was already imported. Moving XFORMERS_DISABLED=1 to the top of the file before any imports to ensure it's set before xformers library initialization. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>13e3d2b -
Fix Stella models CPU compatibility by disabling xformers
Patryk Ptasiński +1: ClaudeStella models were failing on CPU with xformers attention operator errors. Added environment variable XFORMERS_DISABLED=1 for Stella models on CPU to force fallback to standard PyTorch attention mechanisms. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>1a9d56e -
Add cuda
Patryk Ptasiński3726350 -
Add xformers dependency to fix Stella model loading
Patryk Ptasiński +1: Claude🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>1be0f7d -
Fix Stella models to use trust_remote_code=True
Patryk Ptasiński- dunzhang/stella_en_400M_v5 requires custom code execution - dunzhang/stella_en_1.5B_v5 requires custom code execution - infgrad/stella-base-en-v2 requires custom code execution3f9eaee -
Fix nvidia/NV-Embed-v2 to use trust_remote_code=True
Patryk Ptasiński0b5e578 -
Implement single model loading to optimize memory usage
Patryk Ptasiński +1: Claude- Modified load_model function to keep only one model in memory at a time - Previous model is automatically unloaded when loading a new model - Maintains same model reuse optimization if requesting already loaded model - Reduces memory footprint by avoiding accumulation of multiple models 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>3d96b43 -
Add top MTEB leaderboard models for comprehensive coverage
Patryk Ptasiński- Add nvidia/NV-Embed-v2 (ranked #1 on MTEB with 72.31 score) - Add Alibaba-NLP/gte-Qwen2-7B-instruct (ranked #1 in English/Chinese) - Add Alibaba-NLP/gte-Qwen2-1.5B-instruct (efficient 1.5B variant) - Add intfloat/multilingual-e5-large-instruct and multilingual-e5-large - Add BAAI/bge-en-icl (BGE family top performer) - Update model count to 28+ with MTEB top performers highlighted8dc5d16 -
Add Qwen3 and Stella embedding models to predefined list
Patryk Ptasiński- Add Qwen3-Embedding-0.6B, 4B, and 8B models - Add Stella models: dunzhang/stella_en_400M_v5, stella_en_1.5B_v5, infgrad/stella-base-en-v2 - Update model count to 22+ in description1ea76a1 -
Allow any Hugging Face model in API with trust_remote_code=False for unlisted models
Patryk Ptasińskieca0fa7 -
Add fixed height CSS for embedding output panel to prevent excessive page height
Patryk Ptasiński7d3e0ef -
Add snowflake-arctic-embed-m-long model with trust_remote_code=True
Patryk Ptasiński73c7e87 -
Add 15+ embedding models with dropdown selector and comprehensive API support
Patryk Ptasińskib366822 -
Add FastAPI endpoint for direct HTTP access without queue
Patryk Ptasińskicc86f1b -
Update API documentation to clarify Hugging Face Spaces queue requirements
Patryk Ptasiński06197a9 -
Add CLAUDE.md with repository guidance for Claude Code
Patryk Ptasiński3a1e953 -
Disable queue to allow direct API calls
Patryk Ptasiński65ff348 -
Update API documentation to clarify queue system requirements
Patryk Ptasiński9530053 -
Update API documentation with correct endpoints and Python example
Patryk Ptasiński329e33e -
Fix API endpoint by adding submit button with api_name='predict'
Patryk Ptasiński4ac731c -
Add cURL usage guide to Gradio interface
Patryk Ptasińskif336786 -
Convert from GPU to CPU deployment
Patryk Ptasiński- Remove @spaces.GPU decorator - Change device from 'cuda' to 'cpu' - Add explicit CPU-only PyTorch dependencybb5fa37
-
Update requirements.txt
Verified signaturea9f0b60
-
Update requirements.txt
Verified signature58eb928
-
Updating Gradio
Verified signature2e0a50f
-
Upgrading to 1.5
derek-thomas9079015
-
Updating to single text input
derek-thomas7be8da7 -
Adding einops version
derek-thomasfb7f097 -
Adding einops
derek-thomasc9ee68c -
Init commit
derek-thomasa7eec35 -
initial commit
Verified signature2935fa6