-
Following [lesson3](https://www.kaggle.com/hortonhearsafoo/fast-ai-v3-lesson-3-planet)
**Describe the bug**
The error happens after run the learn.export and calling other learn object
[my not…
-
Importing a paper with papis, version 0.8.2 and:
`papis add --from-doi 10.1145/3184558.3186235`
gives following metadata:
```yml
author: Champin, Pierre-Antoine and Gandon, Fabien and Gandon…
-
### Make sure you are using the *latest* version: run `youtube-dl --version` and ensure your version is *2019.03.18*. If it's not, read [this FAQ entry](https://github.com/ytdl-org/youtube-dl/blob/m…
-
**python version:** 3.6.8
**fastai version:** 1.0.34
I downloaded the notebook from https://github.com/fastai/course-v3/blob/master/nbs/dl1/lesson1-pets.ipynb
I installed jupyter notebook, pytorc…
ghost updated
5 years ago
-
**Describe the bug**
After training a Learner and save it, if I create a new Learner and load the state, an error occurs if I try to keep training it. It seems that there is a mismatch on t…
-
**Describe the bug**
The issue is with the tensor types of input(cudaFloat tensor) and the weights(cudaHalfTensor) while calling the ClassificationInterpretation.from_learner. I looked at the s…
-
Current fastai github master + pytorch nightly:
```
class MyLoss():
def __call__(self, pred, t_dummy, t_score):
_, score = pred
self.mse_loss = mean_squared_error(score, t_…
-
**Description**
After installing with conda and cloning the course v3, running the notebook works and then stops at `learn.fit_one_cycle(4)`
```
-------------------------------------------…
-
Something about the toy gridworlds is broken. It looks like a tabular Q-learner learns the worst-case policy instead of the best case one, despite thoroughly exploring the environment.
For referen…
-
I am trying to add my own tool to the framework working with OWL. As I understand, the training examples are defined by the learning problem and can be found in:
./learningtasks/[task_id]/owl/[lp_id]/…