-
Hello,
When I execute ```torch._export.capture_pre_autograd_graph(model, inputs)``` , The the following code throws an error
```
torch.nn.utils.weight_norm(
ConvTranspose1d(
…
-
## Project Roadmap: Domain-Specific Knowledge Mesh
**1. Project Goals:**
* **Unified Data Management:** Create a system that ingests and manages data from various sources, including files, data…
-
### 🐛 Describe the bug
I followed the example to fine-tune HuggingFace's wav2vec 2.0 for speech recognition, using `torch.compile`, aiming to get faster training. However, I ran into an issue as outl…
-
### 🐛 Describe the bug
torch 2.0.1 column parallel output is not sharded (not sure why the documentation is wrong), which seems to be fixed/matching with documentation in 2.1.0 and 2.2.0 (nightly v…
-
Hey @JamesALeedham, thanks for this repo which certainly saved me a lot of headaches. Previously I am not able to generate auto-summary for some of my deeper classes and link them to a new page for d…
-
Hi, thank you for releasing this wonderful code base.
When I'm trying to combine causal and cross-document mask, like below
```
document_ids = document_ids.unsqueeze(1) # [B, S] -> [B, 1, S]
…
-
[Apache Iceberg](https://iceberg.apache.org/) is an Open Table Format (OTF) that can be used for storing data in object storage with a well defined schema.
The Iceberg spec defines its own data sch…
-
### 🚀 The feature, motivation and pitch
Currently when a tensor is marked dynamic via mark_dynamic API. The range is reflected in the graph via ShapeEnv. However, no indication is present whether …
-
## 🚀 The feature, motivation and pitch
### Consumer: `Optimizer._init_group`
The main driving motivation for this API is `Optimizer._init_group`. It is a function that rearranges the input paramet…
-
### 🐛 Describe the bug
The following pattern shows up in a model from User Empathy Day: https://github.com/suno-ai/bark/blob/f4f32d4cd480dfec1c245d258174bc9bde3c2148/bark/generation.py#L602-L607
`…