Add FasterTokenizer model in experiment #1220
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 5 commits
October 22, 2021 06:12
2. add demo with FasterTokenizer usage in experiment
ZeyuChen
reviewed
Oct 27, 2021
| rm -rf * | ||
|
|
||
| # same with the demo.cc | ||
| DEMO_NAME=demo |
There was a problem hiding this comment.
修改为text_cls_infer
| @@ -0,0 +1,64 @@ | |||
|
|
|||
There was a problem hiding this comment.
或者是ernie_infer。同时后面应该还得区分下句子分类还是序列标注任务
There was a problem hiding this comment.
done， 修改为text_cls_infer
There was a problem hiding this comment.
seq_cls_infer/token_cls_infer可能可以跟类名保持更好一致
| "办理入住手续，节省时间。"}; | ||
|
|
||
| std::vector<float> probs; | ||
| Run(predictor.get(), &data, &probs); |
There was a problem hiding this comment.
如果这个demo就是分类，那就写清楚分类的，和序列标注的分开
There was a problem hiding this comment.
seq_cls_infer/token_cls_infer可能可以跟类名保持更好一致
There was a problem hiding this comment.
Done，已修改为seq_cls_infer
| } | ||
|
|
||
| void Run(Predictor* predictor, | ||
| std::vector<std::string>* input_data, |
There was a problem hiding this comment.
修改为
void Run(Predictor* predictor,
const std::vector<std::string>& input_data,
std::vector<float>* logits,
std::vector<int64_t>* predictions)
|
|
||
| import paddle | ||
| import paddlenlp | ||
| from paddlenlp.experimental import FastSequenceClassificationModel |
paddlenlp/experimental/model.py
Outdated
| return logits | ||
|
|
||
|
|
||
| class FastSequenceClassificationModel(object): |
There was a problem hiding this comment.
FasterModelForSequenceClassification
paddlenlp/ops/strings.py
Outdated
| import paddle.fluid.core as core | ||
|
|
||
| __all__ = ['to_string_tensor', 'to_vocab_tensor'] |
paddlenlp/experimental/model.py
Outdated
| raise ValueError("Unknown name %s. Now %s surports %s" % | ||
| (pretrained_model_name_or_path, cls.__name__, | ||
| list(name_model.keys()))) | ||
|
|
There was a problem hiding this comment.
基于这个类新增to_static接口，屏蔽STRINGS类型对外暴露
paddlenlp/experimental/model.py
Outdated
| else: | ||
| raise ValueError("Unknown name %s. Now %s surports %s" % | ||
| (pretrained_model_name_or_path, cls.__name__, | ||
| list(name_model.keys()))) |
There was a problem hiding this comment.
基于这个类新增to_static接口，屏蔽STRINGS类型对外暴露
paddlenlp/ops/strings.py
Outdated
| @@ -12,10 +12,13 @@ | |||
| # See the License for the specific language governing permissions and | |||
| # limitations under the License. | |||
|
|
|||
| import paddle | |||
added 6 commits
October 27, 2021 17:00
2. rename to_vocab_tensor to to_vocab_buff 3. add to_static() api to FasterModel
2. remove some redudant code
2. suuport from_pretrained() with given a local directory
ZeyuChen
reviewed
Oct 30, 2021
| set(CUDA_LIB "/usr/local/cuda/lib64/" CACHE STRING "CUDA Library") | ||
| else() | ||
| if(CUDA_LIB STREQUAL "") | ||
| set(CUDA_LIB "C:\\Program\ Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v8.0\\lib\\x64") |
There was a problem hiding this comment.
这个命令写死路径可能不一定正确，回头得windows测试验证下
| losses.append(loss.numpy()) | ||
| correct = metric.compute(logits, labels) | ||
| metric.update(correct) | ||
| accu = metric.accumulate() |
|
|
||
|
|
||
| def create_dataloader(dataset, mode='train', batch_size=1): | ||
|
|
| text = '小说是文学的一种样式，一般描写人物故事，塑造多种多样的人物形象，但亦有例外。它是拥有不完整布局、发展及主题的文学作品。而对话是不是具有鲜明的个性，每个人物说的没有独特的语言风格，是衡量小说水准的一个重要标准。与其他文学样式相比，小说的容量较大，它可以细致的展现人物性格和命运，可以表现错综复杂的矛盾冲突，同时还可以描述人物所处的社会生活环境。小说一词，最早见于《庄子·外物》：“饰小说以干县令，其于大达亦远矣。”这里所说的小说，是指琐碎的言谈、小的道理，与现时所说的小说相差甚远。文学中，小说通常指长篇小说、中篇、短篇小说和诗的形式。小说是文学的一种样式，一般描写人物故事，塑造多种多样的人物形象，但亦有例外。它是拥有不完整布局、发展及主题的文学作品。而对话是不是具有鲜明的个性，每个人物说的没有独特的语言风格，是衡量小说水准的一个重要标准。与其他文学样式相比，小说的容量较大，它可以细致的展现人物性格和命运，可以表现错综复杂的矛盾冲突，同时还可以描述人物所处的社会生活环境。小说一词，最早见于《庄子·外物》：“饰小说以干县令，其于大达亦远矣。”这里所说的小说，是指琐碎的言谈、小的道理，与现时所说的小说相差甚远。文学中' | ||
| data = [text[:max_seq_length]] * 100 | ||
|
|
||
| pp_tokenizer = FasterTokenizer(vocab, do_lower_case=False) |
There was a problem hiding this comment.
此处接口是否需要与XXXTokenizer.from_pretrained的API体验打平？以及是否需要
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR types
New features
PR changes
APIs
Description