-
### 🐛 Describe the bug
```
import torch
from diffusers import EulerDiscreteScheduler, StableDiffusionPipeline
model_id = "stabilityai/stable-diffusion-2"
scheduler = EulerDiscreteScheduler.from…
-
The scope of work:
- moving CMC, precision, and recall into separate functions (by the analogue with fnmr@fmr)
- adding docs to these functions
- adding these functions to sphinx documentation
-…
-
For split-like ops the output type is a list of (potentially different shape) tensors which is not able to be handled by the current shape refinement pass. The solution introduced in #820 was to do an…
-
Hello. Thank you very much for this software. As a matter of fact, I have encountered a problem that I was wondering if you could help me with. I have successfully implemented all the steps but when i…
-
# 🐛 Bug
Hi guys and thanks a lot for your work.
The predictive variance in my SGPR use cases is broken, approximately since the new release. This happens in both single output and batch mode, it…
-
### 🐛 Describe the bug
Running into issue during tracing:
```
File "/opt/pytorch/pytorch/torch/_prims/context.py", line 207, in _is_func_unsupported_nvfuser
gm = get_isolated_graphmodule(fun…
-
### 🐛 Describe the bug
The model was a very simple conv + batch norm but the error comes from the batchnorm:
```python
import torch
class NORM(torch.nn.Module):
def __init__(self,):
…
-
### 🐛 Describe the bug
To repro:
```python
from functorch.experimental import functionalize
import torch
def f(x):
log_probs = x.log_softmax(2)
targets = torch.randint(1, 20, (16, 30…
-
Hi, I have problem understanding the "Int8 absmax row-wise + decomposition" entry in Table 1. Does it mean "Absmax LLM.int8() (row-wise + decomp)" ? Because it does not contain the "LLM.int8()" keywor…
-
### 🐛 Describe the bug
Although `torch._dynamo`'s `export(f, *args, aten_graph=False, decomposition_table=None, tracing_mode='real', **kwargs)` does allow `**kwargs` on its public API, the followin…