mannix/
llama3.1-storm

548 1 year ago

Llama-3.1-Storm-8B outperforms both Llama-3.1-8B-Instruct and Hermes-3-Llama-3.1-8B!

tools

Models

View all →

13 models

llama3.1-storm:latest

4.7GB · 128K context window · Text · 1 year ago

llama3.1-storm:q2_k

3.2GB · 128K context window · Text · 1 year ago

llama3.1-storm:q3_k_s

3.7GB · 128K context window · Text · 1 year ago

llama3.1-storm:q3_k_l

4.3GB · 128K context window · Text · 1 year ago

llama3.1-storm:q4_0

4.7GB · 128K context window · Text · 1 year ago

llama3.1-storm:q5_0

5.6GB · 128K context window · Text · 1 year ago

llama3.1-storm:q5_1

6.1GB · 128K context window · Text · 1 year ago

llama3.1-storm:q5_k_s

5.6GB · 128K context window · Text · 1 year ago

llama3.1-storm:q5_k_m

5.7GB · 128K context window · Text · 1 year ago

llama3.1-storm:q6_k

6.6GB · 128K context window · Text · 1 year ago

llama3.1-storm:q8_0

8.5GB · 128K context window · Text · 1 year ago

llama3.1-storm:iq3_xxs

3.3GB · 128K context window · Text · 1 year ago

llama3.1-storm:fp16

16GB · 128K context window · Text · 1 year ago

Readme

  • Quantization from fp32
  • Using i-matrix calibration_datav3.txt

🤗 Announcement Blog: https://huggingface.co/blog/akjindal53244/llama31-storm8b

Llama-3.1-Storm-8B: Improved SLM with Self-Curation + Model Merging

image.png

HuggingFace