atsuki-yamaguchi commited on
Commit
bc240f5
1 Parent(s): 329631e

Upload folder using huggingface_hub

Browse files
Files changed (2) hide show
  1. README.md +4 -40
  2. adapter_config.json +1 -26
README.md CHANGED
@@ -1,46 +1,9 @@
1
  ---
2
  library_name: peft
3
- license: mit
4
- language:
5
- - ar
6
  ---
7
- BLOOM-7B Arabic [LAPT + CLP+]
8
- ===
9
-
10
- ## How to use
11
- ```python
12
- from peft import AutoPeftModelForCausalLM
13
- from transformers import AutoTokenizer
14
-
15
- model = AutoPeftModelForCausalLM.from_pretrained(
16
- "atsuki-yamaguchi/bloom-7b1-clpp-ar"
17
- )
18
-
19
- # w/ GPU
20
- model = AutoPeftModelForCausalLM.from_pretrained(
21
- "atsuki-yamaguchi/bloom-7b1-clpp-ar",
22
- device_map="auto",
23
- load_in_8bit=True,
24
- )
25
- ```
26
-
27
- ## Citation
28
- ```
29
- @article{yamaguchi2024empirical,
30
- title={An Empirical Study on Cross-lingual Vocabulary Adaptation for Efficient Generative {LLM} Inference},
31
- author={Atsuki Yamaguchi and Aline Villavicencio and Nikolaos Aletras},
32
- journal={ArXiv},
33
- year={2024},
34
- volume={abs/2402.10712},
35
- url={https://arxiv.org/abs/2402.10712}
36
- }
37
- ```
38
-
39
- ## Link
40
- For more details, please visit https://github.com/gucci-j/llm-cva
41
 
42
 
43
- ## Training procedure
44
  The following `bitsandbytes` quantization config was used during training:
45
  - quant_method: bitsandbytes
46
  - load_in_8bit: True
@@ -52,6 +15,7 @@ The following `bitsandbytes` quantization config was used during training:
52
  - bnb_4bit_quant_type: fp4
53
  - bnb_4bit_use_double_quant: False
54
  - bnb_4bit_compute_dtype: float32
55
-
56
  ### Framework versions
57
- - PEFT 0.5.0
 
 
 
1
  ---
2
  library_name: peft
 
 
 
3
  ---
4
+ ## Training procedure
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5
 
6
 
 
7
  The following `bitsandbytes` quantization config was used during training:
8
  - quant_method: bitsandbytes
9
  - load_in_8bit: True
 
15
  - bnb_4bit_quant_type: fp4
16
  - bnb_4bit_use_double_quant: False
17
  - bnb_4bit_compute_dtype: float32
 
18
  ### Framework versions
19
+
20
+
21
+ - PEFT 0.5.0
adapter_config.json CHANGED
@@ -1,26 +1 @@
1
- {
2
- "auto_mapping": null,
3
- "base_model_name_or_path": "atsuki-yamaguchi/bloom-7b1-clpp-ar",
4
- "bias": "none",
5
- "fan_in_fan_out": false,
6
- "inference_mode": true,
7
- "init_lora_weights": true,
8
- "layers_pattern": null,
9
- "layers_to_transform": null,
10
- "lora_alpha": 32,
11
- "lora_dropout": 0.05,
12
- "modules_to_save": [
13
- "lm_head",
14
- "word_embeddings"
15
- ],
16
- "peft_type": "LORA",
17
- "r": 8,
18
- "revision": null,
19
- "target_modules": [
20
- "query_key_value",
21
- "dense",
22
- "dense_h_to_4h",
23
- "dense_4h_to_h"
24
- ],
25
- "task_type": "CAUSAL_LM"
26
- }
 
1
+ {"auto_mapping": null, "base_model_name_or_path": "atsuki-yamaguchi/bloom-7b1-clpp-ar", "bias": "none", "fan_in_fan_out": false, "inference_mode": true, "init_lora_weights": true, "layers_pattern": null, "layers_to_transform": null, "lora_alpha": 32, "lora_dropout": 0.05, "modules_to_save": ["lm_head", "word_embeddings"], "peft_type": "LORA", "r": 8, "revision": null, "target_modules": ["query_key_value", "dense", "dense_h_to_4h", "dense_4h_to_h"], "task_type": "CAUSAL_LM"}