-
Why does the KeyError occurs?I can not understand.
Can you help me?Thank you
![](http://ofht327si.bkt.clouddn.com/2017-09-12%2006-50-24%E5%B1%8F%E5%B9%95%E6%88%AA%E5%9B%BE.png)
![](http://ofht327si…
-
Right now the code is just one big file. Break it down into classes. Make it GOT-independent, that is add code to use any corpus of data
vshan updated
7 years ago
-
## Environment
- llm-foundry: latest
## To reproduce
Steps to reproduce the behavior:
1. train a `hf_t5` model
2. download the composer checkpoint
3. try to convert it back to huggingface…
-
### Problem Description
When I use seq2seq model,such as TCN or LSTM or Transformer,they are have two inputs,one is x,another is y.And x's shape is different with y's shape,for example x's shape is (…
-
I just tried running the tests using `sudo python setup.py test` and got the following error, which is seems related to the fact that the `shard_to_cpu` parameter was [removed a while ago](https://git…
-
I went through #118 (which address mode.evaluate()) and related issues. However, I am still having issue with **model.fit()** using dropout. Here is the sample code @dajaj posted. It still gives me th…
-
-
Possible strategies:
* Create near duplicates from _orig_ layer of a normalized dataset, e.g. with [augtxt](https://github.com/ulf1/augtxt)
* "Translate" modern text to historical text. Options:
…
-
> G:\Anaconda2\lib\site-packages\gensim\utils.py:840: UserWarning: detected Windows; aliasing chunkize to chunkize_serial
> warnings.warn("detected Windows; aliasing chunkize to chunkize_serial")
…
-
## 🚀 Feature
As proposed in this [PyTorch forum post](https://discuss.pytorch.org/t/using-batches-for-seq2seq-models/55205): Custom sampler for Sequence-to-Sequence models (but also for sequence cl…