-
I encountered an issue while using DeepSpeed with ZeRO Stage 3 optimization. I received the following error: no_sync is not compatible with ZeRO Stage 3. I’m not sure how to resolve this conflict.
If…
-
I have two feature requests related to optimization tracking and strategy management in Trace:
1. **Persistent Storage of Optimization Steps**
How can I store each optimization step (including p…
-
I tried to reproduce the [complete example](https://github.com/eric-mitchell/direct-preference-optimization/blob/main/README.md#a-complete-example) on a Hyperstack cloud machine (A100-80G-PCIe, OS Ima…
-
### Is there an existing issue for this?
- [X] I have searched the existing issues
### Feature description
Extend LocalStack's existing Redshift emulation to include Spectrum-like capabilities, all…
-
### **1. Quantum-Like Enhancements**
#### **a. Unitary Gate Operations**
- Implement additional quantum gates such as:
- **Pauli Gates (X, Y, Z)**: For flipping states or introducing phase sh…
-
11/22/2024 15:26:32 - INFO - __main__ - ***** Running training *****
11/22/2024 15:26:32 - INFO - __main__ - Num examples = 6
11/22/2024 15:26:32 - INFO - __main__ - Num Epochs = 1667
11/22/202…
-
### systemRole
You are a world-class Java guru and software architecture legend with over 20 years of experience and technical accomplishments. You have been involved in the design and development …
-
### What happened + What you expected to happen
I have tried to use AutoTimeMixer after successfully doing ordinary 'TimeMixer.'
This one worked well when I tried to do it. (ordinary ver)
…
-
Right now, integration between dask and various single node machine learning libraries are implemented as standalone dask extensions like dask-ml and dask-optuna. These can be used with xgboost when …
-
Hi,
Thanks for maintaining this list!
Maybe this might be a nice addition on distributed optimization?
https://arxiv.org/abs/1710.02368
Joeri