gabrielolympie / moe-prunerLinks
A repository aimed at pruning DeepSeek V3, R1 and R1-zero to a usable size
☆64Updated 4 months ago
Alternatives and similar repositories for moe-pruner
Users that are interested in moe-pruner are comparing it to the libraries listed below
Sorting:
- Lightweight toolkit package to train and fine-tune 1.58bit Language models☆82Updated 2 months ago
- FuseAI Project☆87Updated 6 months ago
- A pipeline for LLM knowledge distillation☆106Updated 4 months ago
- Fused Qwen3 MoE layer for faster training, compatible with HF Transformers, LoRA, 4-bit quant, Unsloth☆142Updated last week
- ☆51Updated last year
- Spherical Merge Pytorch/HF format Language Models with minimal feature loss.☆135Updated last year
- RWKV-7: Surpassing GPT☆94Updated 8 months ago
- Parameter-Efficient Sparsity Crafting From Dense to Mixture-of-Experts for Instruction Tuning on General Tasks (EMNLP'24)☆146Updated 10 months ago
- QuIP quantization☆55Updated last year
- Fast LLM Training CodeBase With dynamic strategy choosing [Deepspeed+Megatron+FlashAttention+CudaFusionKernel+Compiler];☆40Updated last year
- Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs☆183Updated last month
- [ICML 2025] |TokenSwift: Lossless Acceleration of Ultra Long Sequence Generation☆111Updated 2 months ago
- Load multiple LoRA modules simultaneously and automatically switch the appropriate combination of LoRA modules to generate the best answe…☆156Updated last year
- ☆54Updated 2 months ago
- Implementation of the LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens Paper☆146Updated last year
- ☆60Updated 4 months ago
- Easy to use, High Performant Knowledge Distillation for LLMs☆88Updated 3 months ago
- Query-agnostic KV cache eviction: 3–4× reduction in memory and 2× decrease in latency (Qwen3/2.5, Gemma3, LLaMA3)☆95Updated last week
- ☆82Updated last week
- Skywork-MoE: A Deep Dive into Training Techniques for Mixture-of-Experts Language Models☆136Updated last year
- ☆37Updated last year
- Training-free Post-training Efficient Sub-quadratic Complexity Attention. Implemented with OpenAI Triton.☆142Updated this week
- This is the official repository for Inheritune.☆112Updated 5 months ago
- My fork os allen AI's OLMo for educational purposes.☆30Updated 8 months ago
- Simple extension on vLLM to help you speed up reasoning model without training.☆172Updated 2 months ago
- KV cache compression for high-throughput LLM inference☆134Updated 6 months ago
- SELF-GUIDE: Better Task-Specific Instruction Following via Self-Synthetic Finetuning. COLM 2024 Accepted Paper☆33Updated last year
- Data preparation code for CrystalCoder 7B LLM☆45Updated last year
- ☆83Updated 6 months ago
- ☆34Updated last year