File size: 740 Bytes
eb51aef
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
# Core dependencies for Fire Evacuation RAG System
numpy
torch
transformers
sentence-transformers
gradio

# FAISS for vector similarity search
# Use faiss-cpu for CPU-only systems, or faiss-gpu for GPU systems
faiss-cpu
# faiss-gpu>=1.7.4  # Uncomment if you have CUDA-capable GPU

# Optional: For faster model loading and inference
unsloth  # Faster model loading with Unsloth

# Optional: For model quantization (4-bit/8-bit)
bitsandbytes  # Required for 4-bit/8-bit quantization

# Optional: For optimized attention (FlashAttention2)
# flash-attn>=2.0.0  # Uncomment if you want FlashAttention2 support
# Note: flash-attn requires CUDA and may need to be installed separately
# Install with: pip install flash-attn --no-build-isolation