GPTQ LLM Leaderboard Report #3
Just a few days ago, LLaMA 2 was released by Meta and boy did it make another huge spark in the open-source LLM landscape. I have tested a bunch of LLaMA 2 quantized models by TheBloke and the older LLaMA 1 models for comparison. I will also include more LLaMA2 based fine-tuned models in the next report.
Models Tested
- TheBloke/LLaMA-7b-GPTQ (main)
- TheBloke/LLaMA-7b-GPTQ (gptq-4bit-32g-actorder_True)
- TheBloke/LLaMA-7b-GPTQ (gptq-8bit-64g-actorder_True)
- TheBloke/LLaMA-13b-GPTQ (main)
- TheBloke/LLaMA-13b-GPTQ (gptq-8bit-64g-actorder_True)
- TheBloke/Llama-2-7b-Chat-GPTQ (main)
- TheBloke/Llama-2-7b-Chat-GPTQ (gptq-4bit-32g-actorder_True)
- TheBloke/Llama-2-7B-GPTQ (main)
- TheBloke/Llama-2-7B-GPTQ (gptq-4bit-32g-actorder_True)
- TheBloke/Llama-2-13B-chat-GPTQ (main)
- TheBloke/Llama-2-13B-chat-GPTQ (gptq-8bit-64g-actorder_True)
- TheBloke/Llama-2-13B-GPTQ (main)
- TheBloke/Llama-2-13B-GPTQ (gptq-4bit-32g-actorder_True)
- TheBloke/Llama-2-13B-GPTQ (gptq-8bit-64g-actorder_True)
- TheBloke/Redmond-Puffin-13B-GPTQ (main)
- TheBloke/Redmond-Puffin-13B-GPTQ (gptq-4bit-32g-actorder_True)
- TheBloke/Redmond-Puffin-13B-GPTQ (gptq-8bit-64g-actorder_True)
- TheBloke/airoboros-13B-gpt4-1.4-GPTQ (main)
- TheBloke/airoboros-13B-gpt4-1.4-GPTQ (gptq-4bit-32g-actorder_True)
- TheBloke/airoboros-13B-gpt4-1.4-GPTQ (gptq-8bit-128g-actorder_False)
- TheBloke/upstage-llama-30b-instruct-2048-GPTQ (main)
- TheBloke/airoboros-33B-gpt4-1.4-GPTQ (main)
Results
You can see all the results since the beginning here.