-
Platforms: linux
This test was disabled because it is failing on main branch ([recent examples](https://torch-ci.com/failure?failureCaptures=%5B%22inductor%2Ftest_cpu_cpp_wrapper.py%3A%3ADynamicSha…
-
### System Info
```Shell
- `Accelerate` version: 1.1.0
- Platform: Linux-6.8.0-45-generic-x86_64-with-glibc2.35
- `accelerate` bash location: .venv/bin/accelerate
- Python version: 3.11.10
- N…
-
**Describe the bug**
Megatron-LM doesn't compatible with transformer-engine 1.13.
in transformer-engine:
https://github.com/NVIDIA/TransformerEngine/blob/2643ba1df43397cc84c9da5fe719a66d87ad9a0a/tr…
-
PR #660 contains new IPEX benchmark code. However I see a regression in ipex model loading when testing using larger model such as Qwen 2.5 Coder 32B quantized..
sample code below.. The model neve…
-
I am attempting to emit pytorch code but unfortunately it does not work for fp8, bf16, and int8. I have tried to patch the converter type dict https://github.com/OrenLeung/cutlass/commit/6d619c964eb8b…
-
**Describe the bug**
I’m experiencing an issue when fine-tuning the Llama-2-7b model from Hugging Face with Zero optimization enabled. I am running on 8 Intel Max 1550 GPUs using the code from the exa…
-
I-JEPA is compatible with timm since https://github.com/lightly-ai/lightly/pull/1612
We should update the I-JEPA examples to use timm instead of torchvision:
- [ ] Update `examples/pytorch/ijepa…
-
### Feature request
### Feature request
Enable SmolLM to ["Export to ExecuTorch"](https://github.com/huggingface/transformers/issues/32253) workflow.
#### Instructions
Instructions of how to…
-
This test was disabled because it is failing on main branch ([recent examples](https://torch-ci.com/failure?failureCaptures=%5B%22distributed%2Ftest_c10d_ucc.py%3A%3ADistributedDataParallelTest%3A%3At…
-
## Bug Description
I am running a distributed Linear model (20 parameters) across 2 GPU Nodes, each node having 2 NVIDIA H100 NVL GPUs. The Model uses DDP parallelization strategy. I am generating…