-
As we noted in our email communications, we think the fairest comparisons to make between two graph frameworks are those that offer the best available performance for each at the time the comparisons …
-
Setting `shrinking=True` in `CDClassifier` with loss `log` and penalty `l1` seems to not converge toward the optimal solution. Increasing the number of iteration does not change this.
It looks like…
-
I've been getting inconsistent results with the Deming regression example given in your text. I'm running Tensorflow locally on an iMac using MacOS Sierra 10.12.4. I then explored using different co…
-
Hi Roozbeh - Any update on being able to use block CV spatial blocking outputs with the caret traincontrol function? Cheers
-
Functionality to support basic integration with standard psychometric tools
- mirt/other standard irt tools in r
- lme4 (should be very straightforward; maybe we can show replication of results fr…
-
Hi!
I'm trying to run `cryodrgn abinit_het` on 200K particles downsampled to 128, but the process quickly terminates with an error: `torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to alloca…
-
### Issue Description
Hi, Inference is really slow ~1-2seconds/iteration with `GlobalNoUTurnSampler` with `adapt_step_size=True`. If I turn it off, then inference is much faster. I don't see simila…
-
Hi, @junyanz @taesungp
does anyone encounter issues like the cycle_A loss suddenly becomes very high after a random iteration? my loss log shows that in epoch 78, the discrimintor loss suddenly de…
-
Hello,
I am trying to implement a new compression algorithm for horovod.
I tested it to compress and decompress tensors of different sizes and dimensions, and it works perfectly fine.
My issue i…
-
The PAM/k-medoids implementation appears to implement SWAP, but not the BUILD part for initializing PAM. Instead you have to provide good starting medoids.
I tired benchmarking it on a larger data …
kno10 updated
3 years ago