Update README.md
Browse filesUpdate generations after major fix: https://github.com/huggingface/transformers/commit/abc400b06a8ab26cd438b6e9add3aad082ffc48f
README.md
CHANGED
@@ -63,14 +63,14 @@ It is recommended to directly call the [`generate`](https://huggingface.co/docs/
|
|
63 |
>>> # the fast tokenizer currently does not work correctly
|
64 |
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
|
65 |
|
66 |
-
>>> prompt = "Hello, I'm
|
67 |
|
68 |
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
|
69 |
|
70 |
>>> generated_ids = model.generate(input_ids)
|
71 |
|
72 |
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
|
73 |
-
[
|
74 |
```
|
75 |
|
76 |
By default, generation is deterministic. In order to use the top-k sampling, please set `do_sample` to `True`.
|
@@ -84,7 +84,7 @@ By default, generation is deterministic. In order to use the top-k sampling, ple
|
|
84 |
>>> # the fast tokenizer currently does not work correctly
|
85 |
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
|
86 |
|
87 |
-
>>> prompt = "Hello, I'm
|
88 |
|
89 |
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
|
90 |
|
@@ -92,7 +92,7 @@ By default, generation is deterministic. In order to use the top-k sampling, ple
|
|
92 |
>>> generated_ids = model.generate(input_ids, do_sample=True)
|
93 |
|
94 |
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
|
95 |
-
[
|
96 |
```
|
97 |
|
98 |
### Limitations and bias
|
63 |
>>> # the fast tokenizer currently does not work correctly
|
64 |
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
|
65 |
|
66 |
+
>>> prompt = "Hello, I'm conscious and"
|
67 |
|
68 |
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
|
69 |
|
70 |
>>> generated_ids = model.generate(input_ids)
|
71 |
|
72 |
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
|
73 |
+
['Hello, I am conscious and aware of my surroundings.\nI am also conscious and aware of']
|
74 |
```
|
75 |
|
76 |
By default, generation is deterministic. In order to use the top-k sampling, please set `do_sample` to `True`.
|
84 |
>>> # the fast tokenizer currently does not work correctly
|
85 |
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
|
86 |
|
87 |
+
>>> prompt = "Hello, I'm conscious and"
|
88 |
|
89 |
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
|
90 |
|
92 |
>>> generated_ids = model.generate(input_ids, do_sample=True)
|
93 |
|
94 |
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
|
95 |
+
['Hello, I am conscious and aware.\nSo that makes you dead\nNot always.']
|
96 |
```
|
97 |
|
98 |
### Limitations and bias
|