Instructions to use aufklarer/PersonaPlex-7B-MLX-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use aufklarer/PersonaPlex-7B-MLX-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir PersonaPlex-7B-MLX-4bit aufklarer/PersonaPlex-7B-MLX-4bit
- Moshi
How to use aufklarer/PersonaPlex-7B-MLX-4bit with Moshi:
# pip install moshi # Run the interactive web server python -m moshi.server --hf-repo "aufklarer/PersonaPlex-7B-MLX-4bit" # Then open https://localhost:8998 in your browser
# pip install moshi import torch from moshi.models import loaders # Load checkpoint info from HuggingFace checkpoint = loaders.CheckpointInfo.from_hf_repo("aufklarer/PersonaPlex-7B-MLX-4bit") # Load the Mimi audio codec mimi = checkpoint.get_mimi(device="cuda") mimi.set_num_codebooks(8) # Encode audio (24kHz, mono) wav = torch.randn(1, 1, 24000 * 10) # [batch, channels, samples] with torch.no_grad(): codes = mimi.encode(wav.cuda()) decoded = mimi.decode(codes) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
Quantize depformer to int4 (2.4GB -> 698MB), change default prompt to assistant
Browse files- config.json +2 -1
- depformer.safetensors +2 -2
config.json
CHANGED
|
@@ -69,7 +69,8 @@
|
|
| 69 |
"bits": 4,
|
| 70 |
"group_size": 64,
|
| 71 |
"quantized_components": [
|
| 72 |
-
"temporal"
|
|
|
|
| 73 |
]
|
| 74 |
},
|
| 75 |
"voices": [
|
|
|
|
| 69 |
"bits": 4,
|
| 70 |
"group_size": 64,
|
| 71 |
"quantized_components": [
|
| 72 |
+
"temporal",
|
| 73 |
+
"depformer"
|
| 74 |
]
|
| 75 |
},
|
| 76 |
"voices": [
|
depformer.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf791052e2b953130c1122e6373cde7d2a459162b82d7eea9886351dde4fe0b9
|
| 3 |
+
size 731478800
|