Qwen Llamafiles is an open-source language model by SalmanHabeeb. Features: 619.6m LLM, VRAM: 1.2GB, Context: 32K, License: other, Quantized, LLM Explorer Score: 0.11.
For quantized models, use the GPTQ, AWQ, and GGUF correspondents: 'Qwen1.5-0.5B-Chat-GPTQ-Int4', 'Qwen1.5-0.5B-Chat-GPTQ-Int8', 'Qwen1.5-0.5B-Chat-AWQ', 'Qwen1.5-0.5B-Chat-GGUF'
Supported Languages
English (Proficient), Multilingual (Proficient)
Training Details
Data Volume:
Large amount of data
Methodology:
Transformer-based decoder-only with SwiGLU activation, attention QKV bias, group query attention, post-training with supervised finetuning and direct preference optimization
Context Length:
32768
Model Architecture:
Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, mixture of sliding window attention and full attention
Input Output
Input Format:
Chat message format with role and content
Accepted Modalities:
text
Output Format:
Generated text/content
Performance Tips:
Use provided hyper-parameters in `generation_config.json` for optimal performance