ajayvohra2005
ajayvohra2005
I am unable to share a sample bag but I can share in detail what I am doing. Steps outline: * Converting using ```cv2_to_imgmsg``` with pass through encoding * I...
After trying around different combinations of configurations and vscode-hack extension, the following combination of the vscode-hack version and configuration works, so it appears this issue may have started with vscode-hack...
I am seeing the same issue with DeepSpeed version 0.9.2: ``` self.mpu.get_tensor_model_parallel_group() engine = DeepSpeedHybridEngine(args=args, AttributeError: File "/home/ubuntu/miniconda3/envs/pytorch/lib/python3.10/site-packages/deepspeed/runtime/hybrid_engine.py", line 52, in __init__ 'NoneType' object has no attribute 'get_tensor_model_parallel_group' self.create_inference_module() File...
This PR is specific to Block Manager V2.
> I think the motivation for the proposed change is that in scheduler, 1/ we pad with `0` in block_tables, and 2/ **recompute** when we run out of KV cache...