Commit
•
338af45
1
Parent(s):
8f9fe9b
Update README.md
Browse files
README.md
CHANGED
|
@@ -55,7 +55,25 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
|
|
| 55 |
tokenizer = AutoTokenizer.from_pretrained("AlexWortega/instruct_rugptMedium")
|
| 56 |
model = AutoModelForCausalLM.from_pretrained("AlexWortega/instruct_rugptMedium")
|
| 57 |
```
|
| 58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
# License
|
| 60 |
|
| 61 |
The weights of Instructions ruGPT Small v0.1a are licensed under version 2.0 of the Apache License.
|
|
|
|
| 55 |
tokenizer = AutoTokenizer.from_pretrained("AlexWortega/instruct_rugptMedium")
|
| 56 |
model = AutoModelForCausalLM.from_pretrained("AlexWortega/instruct_rugptMedium")
|
| 57 |
```
|
| 58 |
+
обратите внимание, что лучшие параметры для генерации
|
| 59 |
+
```
|
| 60 |
+
gen_kwargs = {
|
| 61 |
+
"min_length": 20,
|
| 62 |
+
"max_new_tokens": 100,
|
| 63 |
+
"top_k": 50,
|
| 64 |
+
"top_p": 0.9,
|
| 65 |
+
"do_sample": True,
|
| 66 |
+
"early_stopping": True,
|
| 67 |
+
"no_repeat_ngram_size": 2,
|
| 68 |
+
"eos_token_id": tokenizer.eos_token_id,
|
| 69 |
+
"pad_token_id": tokenizer.eos_token_id,
|
| 70 |
+
"use_cache": True,
|
| 71 |
+
"repetition_penalty": 1.5,
|
| 72 |
+
"length_penalty": 0.8,
|
| 73 |
+
"num_beams": 4,
|
| 74 |
+
"num_return_sequences": k
|
| 75 |
+
}
|
| 76 |
+
```
|
| 77 |
# License
|
| 78 |
|
| 79 |
The weights of Instructions ruGPT Small v0.1a are licensed under version 2.0 of the Apache License.
|