-
When fine-tuning the 70b model, I always run into an error while loading the model. Usually, after loading 4 to 10 shards (totally15 shards), the following error occurs(see Error Message).
I'm using…
-
### 🐛 Describe the bug
Running on the latest `viable/strict`, calling just
```
import torch
torch.cuda.memory._record_memory_history(
True, trace_alloc_max_entries=100000, trace_alloc_reco…
awgu updated
2 months ago
-
Hi, great work of this wonderful project! However, I'm having some problems in running cityrefer. I sincerely hope you can help me.
I configured the environment of cityrefer as described in Installa…
-
### 🐛 Describe the bug
### Description
Thanks for the work on this repo.
I am trying to install the nightlies of pytorch, torchvision, and torchtext with pip. Pip downloads several packages wi…
-
### 🐛 Describe the bug
This issue is separated from https://github.com/pytorch/pytorch/issues/104952, verified with TORCHINDUCTOR_FREEZING=0
name
batch_size_new
spe…
-
I ran the example given
```
import os
os.environ["WANDB_DISABLED"] = "true"
!python examples/nlg-reddit/sample-level-dp/fine-tune-dp.py \
--output_dir scratch \
--model_name sshleifer/tiny-gpt2 …
-
I encountered this error when trying to run Whisper([https://github.com/openai/whisper](https://github.com/openai/whisper)) using torch-directml.
Sample code:
```python
import torch_directml
impor…
-
### Your current environment
```text
The output of `python collect_env.py`
```
Collecting environment information...
PyTorch version: 2.4.0+cu121
Is debug build: False
CUDA used to build PyTo…
-
### 🐛 Describe the bug
```python
tempHeteroDict = {};
for key in sageLayerNameList:
tempHeteroDict[key] = GCNConv(sage_dim_in, sage_dim_in);
self.hetero_conv = HeteroConv(tempHeteroDict, aggr…
-
### 🐛 Describe the bug
float32 dynamic shape cpp wrapper
suite
name
thread
batch_size_new
speed_up_new
inductor_new
eager_new
compila…