mqyqlx
commited on
Commit
•
70b9026
1
Parent(s):
83ee88b
update readme
Browse files- README.md +9 -0
- generation_demo.py +1 -1
README.md
CHANGED
@@ -13,6 +13,15 @@ and increases the expressive power of the model by dynamically composing attenti
|
|
13 |
|
14 |
We recommend <strong>compiled version</strong> of DCFormer with *torch.compile* for inference acceleration. Please refer to QuickStart section for compile implementation.
|
15 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
## Quickstart
|
17 |
|
18 |
```
|
|
|
13 |
|
14 |
We recommend <strong>compiled version</strong> of DCFormer with *torch.compile* for inference acceleration. Please refer to QuickStart section for compile implementation.
|
15 |
|
16 |
+
## Env
|
17 |
+
|
18 |
+
You need to upgrade transformers to avoid [(loading problems)](https://github.com/huggingface/transformers/pull/29175).
|
19 |
+
|
20 |
+
```
|
21 |
+
pip install transformers>=4.40.2
|
22 |
+
```
|
23 |
+
|
24 |
+
|
25 |
## Quickstart
|
26 |
|
27 |
```
|
generation_demo.py
CHANGED
@@ -7,7 +7,7 @@ os.environ['TOKENIZERS_PARALLELISM'] = 'false'
|
|
7 |
tokenizer = AutoTokenizer.from_pretrained("Caiyun-AI/DCFormer-2.8B")
|
8 |
model = AutoModelForCausalLM.from_pretrained("Caiyun-AI/DCFormer-2.8B", trust_remote_code=True)
|
9 |
|
10 |
-
device = torch.device('cuda
|
11 |
MAX_BATCH_SIZE = 1
|
12 |
MAX_SEQ_LENGTH = 2048
|
13 |
NUM_TOKENS_TO_GENERATE = 100
|
|
|
7 |
tokenizer = AutoTokenizer.from_pretrained("Caiyun-AI/DCFormer-2.8B")
|
8 |
model = AutoModelForCausalLM.from_pretrained("Caiyun-AI/DCFormer-2.8B", trust_remote_code=True)
|
9 |
|
10 |
+
device = torch.device('cuda')
|
11 |
MAX_BATCH_SIZE = 1
|
12 |
MAX_SEQ_LENGTH = 2048
|
13 |
NUM_TOKENS_TO_GENERATE = 100
|