-
Hi, I am using deepspeed zero3 to fine tune flux model using the script - flux_train_network.py.
```
flux = accelerator.unwrap_model(flux)
print(f"flux - {flux.state_dict()['single_blocks.7.linea…
-
**Description** Unable to launch the FEDOT framework using the `automlbenchmark` repository in Linux. The process fails despite sufficient resources and prior troubleshooting efforts. Works fine on Wi…
-
Good news, zero-config support for Hydrogen has just landed! https://twitter.com/vercel_changes/status/1545515100097970176
-
Currently, we need to provide the EKS cluster's `api-server` and `cluster-certificate` in Bottlerocket's `settings.kubernetes` or else a Bottlerocket instance will not know how to join the EKS cluster…
-
I edited my `/boot/firmware/config.txt` and `/boot/firmware/cmdline.txt` how I always did, with dtoverlay=dwc2 and modules-load=dwc2,g_ether. Than I add a usb0-cfg configuration file to `/etc/network/…
-
Here is an example traceback of me trying to use `heavyball` in TorchTitan. It seems that `heavyball` is not yet compatible with FSDP2 due to some of the utilities operating on tensors instead of dten…
-
I see the config in another issue, and i have a try, but..
> Hi @hpichlerbio, thanks for your interest. One way to finetune on your custom **image** only dataset is to follow the last [section](https…
-
Having this as part of my config:
```
IdleMachines uint64 `yaml:"idle_machines" default:"2"`
```
and in config.yaml if it's
```
idle_machines: 3
```
It has effect
But in config.yaml if it…
-
### 🐛 Describe the bug
```
import torch
import torch._inductor.config
torch.set_default_device("cuda")
inp = torch.empty_strided([4, 512, 96, 96], (1, 4, 196608, 2048), device="cuda")
args…
-
### Confirmation
- [X] This is a bug with an existing resource and is not a feature request or enhancement. Feature requests should be submitted with Cloudflare Support or your account team.
- [X]…