Update README.md
Browse files
README.md
CHANGED
|
@@ -7,20 +7,9 @@ tags:
|
|
| 7 |
- llm
|
| 8 |
- ggml
|
| 9 |
---
|
| 10 |
-
|
| 11 |
-
|
| 12 |
# GGML converted versions of [Mosaic's](https://huggingface.co/mosaicml) MPT Models
|
| 13 |
|
| 14 |
-
|
| 15 |
-
## CAUTION: MPT Development is still ongoing and not finished!
|
| 16 |
-
- Rust & Python: Rustformers implementation see here: [Implement MPT Model](https://github.com/rustformers/llm/pull/218)
|
| 17 |
-
|
| 18 |
-
If these implementations are complete i will add instructions on how to run the models and update them if necesary!
|
| 19 |
-
|
| 20 |
-
|
| 21 |
## Converted Models:
|
| 22 |
-
|
| 23 |
-
|
| 24 |
| Name | Based on | Type | Container |
|
| 25 |
|-|-|-|-|
|
| 26 |
| [mpt-7b-f16.bin](https://huggingface.co/LLukas22/mpt-7b-ggml/blob/main/mpt-7b-f16.bin) | [mpt-7b](https://huggingface.co/mosaicml/mpt-7b) | fp16 | GGML |
|
|
@@ -36,16 +25,46 @@ If these implementations are complete i will add instructions on how to run the
|
|
| 36 |
| [mpt-7b-storywriter-q4_0.bin](https://huggingface.co/LLukas22/mpt-7b-ggml/blob/main/mpt-7b-storywriter-q4_0.bin) | [mpt-7b-storywriter](https://huggingface.co/mosaicml/mpt-7b-storywriter) | int4 | GGML |
|
| 37 |
| [mpt-7b-storywriter-q4_0-ggjt.bin](https://huggingface.co/LLukas22/mpt-7b-ggml/blob/main/mpt-7b-storywriter-q4_0-ggjt.bin) | [mpt-7b-storywriter](https://huggingface.co/mosaicml/mpt-7b-storywriter) | int4 | GGJT |
|
| 38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
|
| 40 |
-
|
|
|
|
|
|
|
| 41 |
|
| 42 |
-
|
| 43 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
|
| 45 |
-
###
|
| 46 |
The `GGML` example only supports the ggml container type!
|
| 47 |
|
| 48 |
-
|
| 49 |
|
| 50 |
```
|
| 51 |
git clone https://github.com/ggerganov/ggml
|
|
@@ -55,7 +74,7 @@ cmake ..
|
|
| 55 |
make -j4 mpt
|
| 56 |
```
|
| 57 |
|
| 58 |
-
|
| 59 |
|
| 60 |
```
|
| 61 |
./bin/mpt -m path/to/model.bin -p "The meaning of life is"
|
|
|
|
| 7 |
- llm
|
| 8 |
- ggml
|
| 9 |
---
|
|
|
|
|
|
|
| 10 |
# GGML converted versions of [Mosaic's](https://huggingface.co/mosaicml) MPT Models
|
| 11 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
## Converted Models:
|
|
|
|
|
|
|
| 13 |
| Name | Based on | Type | Container |
|
| 14 |
|-|-|-|-|
|
| 15 |
| [mpt-7b-f16.bin](https://huggingface.co/LLukas22/mpt-7b-ggml/blob/main/mpt-7b-f16.bin) | [mpt-7b](https://huggingface.co/mosaicml/mpt-7b) | fp16 | GGML |
|
|
|
|
| 25 |
| [mpt-7b-storywriter-q4_0.bin](https://huggingface.co/LLukas22/mpt-7b-ggml/blob/main/mpt-7b-storywriter-q4_0.bin) | [mpt-7b-storywriter](https://huggingface.co/mosaicml/mpt-7b-storywriter) | int4 | GGML |
|
| 26 |
| [mpt-7b-storywriter-q4_0-ggjt.bin](https://huggingface.co/LLukas22/mpt-7b-ggml/blob/main/mpt-7b-storywriter-q4_0-ggjt.bin) | [mpt-7b-storywriter](https://huggingface.co/mosaicml/mpt-7b-storywriter) | int4 | GGJT |
|
| 27 |
|
| 28 |
+
⚠️Caution⚠️: mpt-7b-storywriter is still under development!
|
| 29 |
+
|
| 30 |
+
## Usage
|
| 31 |
+
|
| 32 |
+
### Python via [llm-rs](https://github.com/LLukas22/llm-rs-python):
|
| 33 |
+
|
| 34 |
+
#### Installation
|
| 35 |
+
Via pip: `pip install llm-rs huggingface_hub`
|
| 36 |
|
| 37 |
+
#### Run inference
|
| 38 |
+
```python
|
| 39 |
+
from llm_rs import Mpt
|
| 40 |
|
| 41 |
+
#Download the model
|
| 42 |
+
hf_hub_download(repo_id="LLukas22/mpt-7b-ggml", filename="mpt-7b-q4_0-ggjt.bin", local_dir=".")
|
| 43 |
+
|
| 44 |
+
#Load the model
|
| 45 |
+
model = Mpt("mpt-7b-q4_0-ggjt.bin")
|
| 46 |
+
|
| 47 |
+
#Generate
|
| 48 |
+
print(model.generate("The meaning of life is"))
|
| 49 |
+
```
|
| 50 |
+
|
| 51 |
+
### Rust via [Rustformers/llm](https://github.com/rustformers/llm):
|
| 52 |
+
|
| 53 |
+
#### Installation
|
| 54 |
+
```
|
| 55 |
+
git clone --recurse-submodules [email protected]:rustformers/llm.git
|
| 56 |
+
cargo build --release
|
| 57 |
+
```
|
| 58 |
+
|
| 59 |
+
#### Run inference
|
| 60 |
+
```
|
| 61 |
+
cargo run --release -- mpt infer -m path/to/model.bin -p "Tell me how cool the Rust programming language is:"
|
| 62 |
+
```
|
| 63 |
|
| 64 |
+
### C via [GGML](https://github.com/ggerganov/ggml)
|
| 65 |
The `GGML` example only supports the ggml container type!
|
| 66 |
|
| 67 |
+
#### Installation
|
| 68 |
|
| 69 |
```
|
| 70 |
git clone https://github.com/ggerganov/ggml
|
|
|
|
| 74 |
make -j4 mpt
|
| 75 |
```
|
| 76 |
|
| 77 |
+
#### Run inference
|
| 78 |
|
| 79 |
```
|
| 80 |
./bin/mpt -m path/to/model.bin -p "The meaning of life is"
|