Skip to content

[Doc] Update doc for prebuilt models #767

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Aug 17, 2023
Merged

[Doc] Update doc for prebuilt models #767

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
41a5954
Update doc for prebuilt models
CharlieFRuan Aug 16, 2023
48f0819
Change model series to original model
CharlieFRuan Aug 17, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
73 changes: 58 additions & 15 deletions docs/prebuilt_models.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,15 +25,17 @@ Prebuilt Models for CLI
:header-rows: 1

* - Model code
- Model Series
- Original Model
- Quantization Mode
- Hugging Face repo
* - `Llama-2-7b-q4f16_1`
- `Llama <https://ai.meta.com/llama/>`__
* - `Llama-2-{7, 13, 70}b-chat-hf-q4f16_1`
- `Llama-2 <https://ai.meta.com/llama/>`__
- * Weight storage data type: int4
* Running data type: float16
* Symmetric quantization
- `link <https://huggingface.co/mlc-ai/mlc-chat-Llama-2-7b-chat-hf-q4f16_1>`__
- * `7B link <https://huggingface.co/mlc-ai/mlc-chat-Llama-2-7b-chat-hf-q4f16_1>`__
* `13B link <https://huggingface.co/mlc-ai/mlc-chat-Llama-2-13b-chat-hf-q4f16_1>`__
* `70B link <https://huggingface.co/mlc-ai/mlc-chat-Llama-2-70b-chat-hf-q4f16_1>`__
* - `vicuna-v1-7b-q3f16_0`
- `Vicuna <https://lmsys.org/blog/2023-03-30-vicuna/>`__
- * Weight storage data type: int3
Expand All @@ -46,24 +48,58 @@ Prebuilt Models for CLI
* Running data type: float16
* Symmetric quantization
- `link <https://huggingface.co/mlc-ai/mlc-chat-RedPajama-INCITE-Chat-3B-v1-q4f16_1>`__
* - `rwkv-raven-1b5-q8f16_0`
* - `rwkv-raven-{1b5, 3b, 7b}-q8f16_0`
- `RWKV <https://github.com/BlinkDL/RWKV-LM>`__
- * Weight storage data type: uint8
* Running data type: float16
* Symmetric quantization
- `link <https://huggingface.co/mlc-ai/mlc-chat-rwkv-raven-1b5-q8f16_0>`__
* - `rwkv-raven-3b-q8f16_0`
- `RWKV <https://github.com/BlinkDL/RWKV-LM>`__
- * Weight storage data type: uint8
* Running data type: float16
- * `1b5 link <https://huggingface.co/mlc-ai/mlc-chat-rwkv-raven-1b5-q8f16_0>`__
* `3b link <https://huggingface.co/mlc-ai/mlc-chat-rwkv-raven-3b-q8f16_0>`__
* `7b link <https://huggingface.co/mlc-ai/mlc-chat-rwkv-raven-7b-q8f16_0>`__
* - `WizardLM-13B-V1.2-{q4f16_1, q4f32_1}`
- `WizardLM <https://github.com/nlpxucan/WizardLM>`__
- * Weight storage data type: int4
* Running data type: float{16, 32}
* Symmetric quantization
- `link <https://huggingface.co/mlc-ai/mlc-chat-rwkv-raven-3b-q8f16_0>`__
* - `rwkv-raven-7b-q8f16_0`
- `RWKV <https://github.com/BlinkDL/RWKV-LM>`__
- * Weight storage data type: uint8
- * `q4f16_1 link <https://huggingface.co/mlc-ai/mlc-chat-WizardLM-13B-V1.2-q4f16_1>`__
* `q4f32_1 link <https://huggingface.co/mlc-ai/mlc-chat-WizardLM-13B-V1.2-q4f32_1>`__
* - `WizardCoder-15B-V1.0-{q4f16_1, q4f32_1}`
- `WizardCoder <https://github.com/nlpxucan/WizardLM>`__
- * Weight storage data type: int4
* Running data type: float{16, 32}
* Symmetric quantization
- * `q4f16_1 link <https://huggingface.co/mlc-ai/mlc-chat-WizardCoder-15B-V1.0-q4f16_1>`__
* `q4f32_1 link <https://huggingface.co/mlc-ai/mlc-chat-WizardCoder-15B-V1.0-q4f32_1>`__
* - `WizardMath-{7, 13, 70}B-V1.0-q4f16_1`
- `WizardMath <https://github.com/nlpxucan/WizardLM>`__
- * Weight storage data type: int4
* Running data type: float16
* Symmetric quantization
- `link <https://huggingface.co/mlc-ai/mlc-chat-rwkv-raven-7b-q8f16_0>`__
- * `7B link <https://huggingface.co/mlc-ai/mlc-chat-WizardMath-7B-V1.0-q4f16_1>`__
* `13B link <https://huggingface.co/mlc-ai/mlc-chat-WizardMath-13B-V1.0-q4f16_1>`__
* `70B link <https://huggingface.co/mlc-ai/mlc-chat-WizardMath-70B-V1.0-q4f16_1>`__
* - `llama2-7b-chat-uncensored-{q4f16_1, q4f32_1}`
- `georgesung <https://huggingface.co/georgesung/llama2_7b_chat_uncensored>`__
- * Weight storage data type: int4
* Running data type: float{16, 32}
* Symmetric quantization
- * `q4f16_1 link <https://huggingface.co/mlc-ai/mlc-chat-georgesung-llama2-7b-chat-uncensored-q4f16_1>`__
* `q4f32_1 link <https://huggingface.co/mlc-ai/mlc-chat-georgesung-llama2-7b-chat-uncensored-q4f32_1>`__
* - `Llama2-Chinese-7b-Chat-{q4f16_1, q4f32_1}`
- `FlagAlpha <https://github.com/FlagAlpha/Llama2-Chinese>`__
- * Weight storage data type: int4
* Running data type: float{16, 32}
* Symmetric quantization
- * `q4f16_1 link <https://huggingface.co/mlc-ai/mlc-chat-FlagAlpha-Llama2-Chinese-7b-Chat-q4f16_1>`__
* `q4f32_1 link <https://huggingface.co/mlc-ai/mlc-chat-FlagAlpha-Llama2-Chinese-7b-Chat-q4f32_1>`__
* - `GOAT-7B-Community-{q4f16_1, q4f32_1}`
- `GOAT-AI <https://huggingface.co/GOAT-AI/GOAT-7B-Community>`__
- * Weight storage data type: int4
* Running data type: float{16, 32}
* Symmetric quantization
- * `q4f16_1 link <https://huggingface.co/mlc-ai/mlc-chat-GOAT-7B-Community-q4f16_1>`__
* `q4f32_1 link <https://huggingface.co/mlc-ai/mlc-chat-GOAT-7B-Community-q4f32_1>`__


To download and run one model with CLI, follow the instructions below:

Expand Down Expand Up @@ -179,6 +215,11 @@ For example, if you compile `OpenLLaMA-7B <https://github.com/openlm-research/op
* - Model library name
- Model Family
- Quantization Mode
* - `Llama-2-7b-chat-hf-q3f16_1`
- LLaMA
- * Weight storage data type: int3
* Running data type: float16
* Symmetric quantization
* - `vicuna-v1-7b-q3f16_0`
- LLaMA
- * Weight storage data type: int3
Expand Down Expand Up @@ -247,6 +288,8 @@ MLC-LLM supports the following model architectures:
* `Gorilla <https://huggingface.co/gorilla-llm/gorilla-7b-hf-delta-v0>`__
* `WizardLM <https://github.com/nlpxucan/WizardLM>`__
* `YuLan-Chat <https://github.com/RUC-GSAI/YuLan-Chat>`__
* `WizardMath <https://github.com/nlpxucan/WizardLM/tree/main/WizardMath>`__
* `FlagAlpha Llama-2 Chinese <https://github.com/FlagAlpha/Llama2-Chinese>`__
* - ``gpt-neox``
- `GPT-NeoX <https://github.com/EleutherAI/gpt-neox>`__
- `Relax Code <https://github.com/mlc-ai/mlc-llm/blob/main/mlc_llm/relax_model/gpt_neox.py>`__
Expand Down