Comments (5)
Hey!
StoppingCriteria was modified in #29116 to return a tensor of shape (batch_size,) where each entry is a boolean indicating if to stop or not. So you can do something like:
eos_token = torch.tensor(50256, device=input_ids.device)
return torch.isin(input_ids[:, -1], eos_token)
from transformers.
transformers==4.38.0
is OK but transformers==4.39.0
exists bug
from transformers.
cc @gante
from transformers.
Thanks very much
from transformers.
Thanks very much
from transformers.
Related Issues (20)
- 4.42.4version bug,@torch._custom_ops.impl_abstract("torchvision::nms") HOT 1
- pipeline gives a different result than the other approach in predicting word probability HOT 2
- DoLa decoding fails on cuda HOT 3
- Training Fails with attn_implementation="flash_attention_2" in gemma-2-9b Model HOT 2
- LLama can't use `torch.compile()` HOT 2
- Add support for Apple's DCLM-Baseline-7B model HOT 3
- GPT-2 Model Logits and Loss are different on MPS HOT 7
- Arrow files too large when training SegFormer HOT 2
- Questions about outputs.logits HOT 3
- ValueError: InternVLChatModel does not support Flash Attention 2.0 yet. HOT 2
- model.generate cannot handle past_key_values correctly HOT 2
- Embedding size 0 when using TrainingArguments & Deepspeed stage 3 with ```model.get_input_embedding()``` HOT 1
- Chameleon model failed after receiving two times the same inputs HOT 2
- Transformer Hangup. HOT 2
- Minor typo in ImageClassificationPipeline
- Bert cannot converge on toy dataset HOT 1
- Why MPS can never be used successfully? HOT 5
- ddp_time in TrainingArguments with deepspeed doesn't take effect HOT 1
- static cache implementation is not compatible with attn_implementation==flash_attention_2 HOT 2
- `AutoModel` class for `image-text-to-text` models HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from transformers.