Upload folder using huggingface_hub
Browse files- README.md +21 -0
- config.json +42 -0
- pytorch_model.bin +3 -0
README.md
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model:
|
3 |
+
- medmekk/Llama-3.2-1B-torchao-int8da8w
|
4 |
+
---
|
5 |
+
|
6 |
+
# medmekk/Llama-3.2-1B-torchao-int8da8w (Quantized)
|
7 |
+
|
8 |
+
## Description
|
9 |
+
This model is a quantized version of the original model `medmekk/Llama-3.2-1B-torchao-int8da8w`. It has been quantized using int8_dynamic_activation_int8_weight quantization with torchao.
|
10 |
+
|
11 |
+
## Quantization Details
|
12 |
+
- **Quantization Type**: int8_dynamic_activation_int8_weight
|
13 |
+
- **Group Size**: None
|
14 |
+
|
15 |
+
## Usage
|
16 |
+
You can use this model in your applications by loading it directly from the Hugging Face Hub:
|
17 |
+
|
18 |
+
```python
|
19 |
+
from transformers import AutoModel
|
20 |
+
|
21 |
+
model = AutoModel.from_pretrained("medmekk/Llama-3.2-1B-torchao-int8da8w")
|
config.json
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "meta-llama/Llama-3.2-1B",
|
3 |
+
"architectures": [
|
4 |
+
"LlamaModel"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 128000,
|
9 |
+
"eos_token_id": 128001,
|
10 |
+
"head_dim": 64,
|
11 |
+
"hidden_act": "silu",
|
12 |
+
"hidden_size": 2048,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 8192,
|
15 |
+
"max_position_embeddings": 131072,
|
16 |
+
"mlp_bias": false,
|
17 |
+
"model_type": "llama",
|
18 |
+
"num_attention_heads": 32,
|
19 |
+
"num_hidden_layers": 16,
|
20 |
+
"num_key_value_heads": 8,
|
21 |
+
"pretraining_tp": 1,
|
22 |
+
"quantization_config": {
|
23 |
+
"modules_to_not_convert": null,
|
24 |
+
"quant_method": "torchao",
|
25 |
+
"quant_type": "int8_dynamic_activation_int8_weight",
|
26 |
+
"quant_type_kwargs": {}
|
27 |
+
},
|
28 |
+
"rms_norm_eps": 1e-05,
|
29 |
+
"rope_scaling": {
|
30 |
+
"factor": 32.0,
|
31 |
+
"high_freq_factor": 4.0,
|
32 |
+
"low_freq_factor": 1.0,
|
33 |
+
"original_max_position_embeddings": 8192,
|
34 |
+
"rope_type": "llama3"
|
35 |
+
},
|
36 |
+
"rope_theta": 500000.0,
|
37 |
+
"tie_word_embeddings": true,
|
38 |
+
"torch_dtype": "bfloat16",
|
39 |
+
"transformers_version": "4.48.0.dev0",
|
40 |
+
"use_cache": true,
|
41 |
+
"vocab_size": 128256
|
42 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:766d64ff8292c2070bc323defba2482acf603178a58b3f99720a227360206580
|
3 |
+
size 1502463122
|