Skip to content

Commit

Permalink
docs: simplify model table
Browse files Browse the repository at this point in the history
  • Loading branch information
fboulnois committed Aug 17, 2024
1 parent de90882 commit a101a5b
Showing 1 changed file with 10 additions and 8 deletions.
18 changes: 10 additions & 8 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -43,12 +43,14 @@ where `<model>` is the name of the model. By default, these will download the
`_Q5_K_M.gguf` versions of the models. These models are quantized to 5 bits
which provide a good balance between speed and accuracy.

| Target | Model | Model Size | (V)RAM Required | [~Score](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) | [~ELO](https://chat.lmsys.org/?leaderboard) | Notes |
Confused about which model to use? Below is a list of popular models, ranked by
[ELO rating](https://en.wikipedia.org/wiki/Elo_rating_system). Generally, the
higher the ELO rating the better the model.

| Target | Model | Parameters | Size | [~Score](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) | [~ELO](https://chat.lmsys.org/?leaderboard) | Notes |
| --- | --- | --- | --- | --- | --- | --- |
| llama-2-13b | [`llama-2-13b-chat`](https://huggingface.co/TheBloke/Llama-2-13B-chat-GGUF) | 13B | 11.73 GB | 55.69 | 1067 | The original open LLM |
| solar-10b | [`solar-10.7b-instruct-v1.0`](https://huggingface.co/TheBloke/SOLAR-10.7B-Instruct-v1.0-GGUF) | 10.7B | 10.10 GB | 74.20 | 1066 | Scores highly on Huggingface |
| phi-3-mini | [`phi-3-mini`](https://huggingface.co/bartowski/Phi-3-mini-4k-instruct-GGUF) | 3B | 3.13 GB | 69.91 | 1074 | The current best tiny model |
| mistral-7b | [`mistral-7b-instruct-v0.2`](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF) | 7B | 7.63 GB | 65.71 | 1074 | The most popular 7B model |
| starling-7b | [`starling-lm-7b-beta`](https://huggingface.co/bartowski/Starling-LM-7B-beta-GGUF) | 7B | 5.13 GB | 69.88 | 1118 | Scores highly on Huggingface |
| command-r | [`c4ai-command-r-v01`](https://huggingface.co/bartowski/c4ai-command-r-v01-GGUF) | 35B | 25.00 GB | 68.54 | 1149 | The current best medium model |
| llama-3-8b | [`meta-llama-3-8b-instruct`](https://huggingface.co/bartowski/Meta-Llama-3-8B-Instruct-GGUF) | 8B | 5.73 GB | 62.55 | 1154 | The current best small model |
| gemma-2-9b | [`gemma-2-9b-it`](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF) | 9B | 6.65 GB | 28.90 | 1187 | Google's best small model |
| llama-3-8b | [`meta-llama-3.1-8b-instruct`](https://huggingface.co/bartowski/Meta-Llama-3.1-8B-Instruct-GGUF) | 8B | 5.73 GB | 26.59 | 1162 | The overall best small model |
| mistral-7b | [`mistral-7b-instruct-v0.2`](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF) | 7B | 5.13 GB | 18.44 | 1072 | The most popular 7B model |
| phi-3-mini | [`phi-3-mini-4k-instruct`](https://huggingface.co/bartowski/Phi-3-mini-4k-instruct-GGUF) | 3B | 3.13 GB | 25.97 | 1066 | The current best tiny model |
| llama-2-13b | [`llama-2-13b-chat`](https://huggingface.co/TheBloke/Llama-2-13B-chat-GGUF) | 13B | 9.23 GB | 11.00 | 1063 | The original open LLM |

0 comments on commit a101a5b

Please sign in to comment.