@snarkai
I run python -m spacy download en
,and I get a “Permission denied” message
I tried to solve it through https://github.com/snarkai/snark-doc, but it didn’t work
@snarkai
I run python -m spacy download en
,and I get a “Permission denied” message
Refer to this issue,
I think it has been solved with Pytorch 0.4
Sudo doesn’t work, neither does pip. Activating the conda environment and then installing also doesn’t work. How do i upgrade pytorch to 0.4?
Hello, it has been working for the snark GPU, but when i start again it gives me this error:
‘File “/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/snark/cli/connect.py”, line 44, in connect
raise NotImplemented()
TypeError: ‘NotImplementedType’ object is not callable’
while the pod is started but does not want to connect local host:8888.
Thank you,
Danny
just to make it clear it used to worked, i have been using it several times.
But now suddenly it did not work.
Hey @diskandar what’s the version of the snark package? Can you try pip3 install --upgrade --no-cache snark
Another possiblity is that you already have something running on port 8888. Run this to check
$sudo netstat -tulpn | grep 8888
Hi thank you now it works with pip3 install --upgrade --no-cache snark
although I have no idea why that thing happen, this is what frustrating about learning this AI.
There are just so many thing that could go wrong on the setting up phase, even before I touch the real AI stuff. @jeremy @rachel
But thank you again!
@diskandar We’re iterating fast and pushing new updates, old code becomes deprecated, our package needs to be updated. Apologies for the inconvenience. We aim to have effortless setup.
pip3 install --upgrade --no-cache snark
snark start --pod_type fast.ai --jupyter
and you can access our cloud GPU servers in your own browser.
Does anyone know the new Promo code for Fast.Ai Learners? I’ve tried at least 4.
FASTAIG5H7
UWQMTGL
FASTAI6GKZ
FASTAI15
That is for Paperspace.
It is in the first post
Hi, I’ve been running lesson4-imdb notebook on p106 instance.
and run into this error
learner.fit(3e-3, 4, wds=1e-6, cycle_len=1, cycle_mult=2)
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
<ipython-input-13-357a8890c905> in <module>()
----> 1 learner.fit(3e-3, 4, wds=1e-6, cycle_len=1, cycle_mult=2)
~/courses/dl1/fastai/learner.py in fit(self, lrs, n_cycle, wds, **kwargs)
302 self.sched = None
303 layer_opt = self.get_layer_opt(lrs, wds)
--> 304 return self.fit_gen(self.model, self.data, layer_opt, n_cycle, **kwargs)
305
306 def warm_up(self, lr, wds=None):
~/courses/dl1/fastai/learner.py in fit_gen(self, model, data, layer_opt, n_cycle, cycle_len, cycle_mult, cycle_save_name, best_save_name, use_clr, use_clr_beta, metrics, callbacks, use_wd_sched, norm_wds, wds_sched_mult, use_swa, swa_start, swa_eval_freq, **kwargs)
249 metrics=metrics, callbacks=callbacks, reg_fn=self.reg_fn, clip=self.clip, fp16=self.fp16,
250 swa_model=self.swa_model if use_swa else None, swa_start=swa_start,
--> 251 swa_eval_freq=swa_eval_freq, **kwargs)
252
253 def get_layer_groups(self): return self.models.get_layer_groups()
~/courses/dl1/fastai/model.py in fit(model, data, n_epochs, opt, crit, metrics, callbacks, stepper, swa_model, swa_start, swa_eval_freq, visualize, **kwargs)
159
160 if not all_val:
--> 161 vals = validate(model_stepper, cur_data.val_dl, metrics, seq_first=seq_first)
162 stop=False
163 for cb in callbacks: stop = stop or cb.on_epoch_end(vals)
~/courses/dl1/fastai/model.py in validate(stepper, dl, metrics, seq_first)
233 for (*x,y) in iter(dl):
234 y = VV(y)
--> 235 preds, l = stepper.evaluate(VV(x), y)
236 batch_cnts.append(batch_sz(x, seq_first=seq_first))
237 loss.append(to_np(l))
~/courses/dl1/fastai/model.py in evaluate(self, xs, y)
77 preds = self.m(*xs)
78 if isinstance(preds,tuple): preds=preds[0]
---> 79 return preds, self.crit(preds, y)
80
81 def set_train_mode(m):
/opt/conda/envs/fastai/lib/python3.6/site-packages/torch/nn/functional.py in cross_entropy(input, target, weight, size_average, ignore_index, reduce)
1159 >>> loss.backward()
1160 """
-> 1161 return nll_loss(log_softmax(input, 1), target, weight, size_average, ignore_index, reduce)
1162
1163
/opt/conda/envs/fastai/lib/python3.6/site-packages/torch/nn/functional.py in log_softmax(input, dim, _stacklevel)
784 if dim is None:
785 dim = _get_softmax_dim('log_softmax', input.dim(), _stacklevel)
--> 786 return torch._C._nn.log_softmax(input, dim)
787
788
RuntimeError: cuda runtime error (2) : out of memory at /opt/conda/conda-bld/pytorch_1518244421288/work/torch/lib/THC/generic/THCStorage.cu:58
I read on the forum that this notebook can cause this kind of error, so I decreased batch size from 64 to 32, and while GPU 1080 were not available started training network again. After 1.5 hours it gave same error. But I lost track if it finished 1 epoch or not. Any suggestions except decreasing batch size?
The error remains same
Installing collected packages: pytorch
Running setup.py install for pytorch ... error
Complete output from command /opt/conda/envs/fastai/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-install-pzov72we/pytorch/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-record-8ylbpu2q/install-record.txt --single-version-externally-managed --compile:
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/tmp/pip-install-pzov72we/pytorch/setup.py", line 13, in <module>
raise Exception(message)
Exception: You should install pytorch from http://pytorch.org
----------------------------------------
Command "/opt/conda/envs/fastai/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-install-pzov72we/pytorch/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-record-8ylbpu2q/install-record.txt --single-version-externally-managed --compile"failed with error code 1 in /tmp/pip-install-pzov72we/pytorch/
I have the same problem
You will need to decrease the batch even more.
But to prevent losing training, you can use best_save_name=‘savedmodel’ in the fit function. With that, fit will save the trained model in a file named savedmodel
when your metric indicator improves.