zhaozitian
commited on
Commit
•
bf456b4
1
Parent(s):
d19b4cb
Update README.md
Browse files
README.md
CHANGED
@@ -1,32 +1,33 @@
|
|
1 |
---
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
4 |
-
## Training procedure
|
5 |
|
|
|
|
|
6 |
|
7 |
-
The
|
8 |
-
|
9 |
-
-
|
10 |
-
- llm_int8_skip_modules: None
|
11 |
-
- llm_int8_enable_fp32_cpu_offload: False
|
12 |
|
13 |
-
|
14 |
-
- load_in_8bit: True
|
15 |
-
- llm_int8_threshold: 6.0
|
16 |
-
- llm_int8_skip_modules: None
|
17 |
-
- llm_int8_enable_fp32_cpu_offload: False
|
18 |
-
### Framework versions
|
19 |
|
20 |
-
|
21 |
|
22 |
-
|
23 |
-
|
|
|
24 |
- load_in_8bit: True
|
25 |
- llm_int8_threshold: 6.0
|
26 |
- llm_int8_skip_modules: None
|
27 |
- llm_int8_enable_fp32_cpu_offload: False
|
|
|
28 |
### Framework versions
|
29 |
|
30 |
- PEFT 0.5.0.dev0
|
31 |
|
32 |
-
|
|
|
1 |
---
|
2 |
+
license: cc-by-sa-4.0
|
3 |
+
datasets:
|
4 |
+
- izumi-lab/llm-japanese-dataset
|
5 |
+
language:
|
6 |
+
- ja
|
7 |
+
- en
|
8 |
---
|
|
|
9 |
|
10 |
+
## This model is a fine-tuned Llama2-7b-chat-hf model with Japanese dataset with LoRA.
|
11 |
+
# This model is finetuned by the joint efforts of Sparticle Inc. and A. I. Hakusan Inc.
|
12 |
|
13 |
+
The training set of this model contains:
|
14 |
+
|
15 |
+
5% of randomly chosen data from llm-japanese-dataset by izumi-lab.
|
|
|
|
|
16 |
|
17 |
+
Japanese-alpaca-lora dataset, retrieved from https://github.com/masa3141/japanese-alpaca-lora/tree/main
|
|
|
|
|
|
|
|
|
|
|
18 |
|
19 |
+
For inference, please follow the instructions in https://github.com/tloen/alpaca-lora/ .
|
20 |
|
21 |
+
## Training procedure
|
22 |
+
|
23 |
+
The following `bitsandbytes` quantization config was used during training:
|
24 |
- load_in_8bit: True
|
25 |
- llm_int8_threshold: 6.0
|
26 |
- llm_int8_skip_modules: None
|
27 |
- llm_int8_enable_fp32_cpu_offload: False
|
28 |
+
|
29 |
### Framework versions
|
30 |
|
31 |
- PEFT 0.5.0.dev0
|
32 |
|
33 |
+
You must agree with Meta's agreements when using this LoRA adapter with Llama-2.
|