NVIDIA/NeMo-Framework-Launcher

Issues

• slurm Multi-machine and multi-GPU training

  #416 opened 4 months ago by yangzhipeng1108
  0

• Adding support for Ray as a launcher backend

  #387 opened 5 months ago by Irvingwangjr
  0

• ’General Configuration‘ Page Not Found

  #373 opened 5 months ago by Pty72
  0

• `demonic child error` occur when execute run_dask_stage.sh

  #340 opened 7 months ago by eagle705
  1

• The hydra-core and omegaconf file encountered type error unexpected keyword vesion base

  #338 opened 7 months ago by imn00b
  1

• [Request][Nemo-Curator] missing add_id and download_and_extract features on Launcher

  #339 opened 7 months ago by leejinho610
  1

• NeMo FW Launcher doesn't have `add_id` and `download_and_extract` logic unlike NeMo Curator repo

  #334 opened 7 months ago by eagle705
  0

• Data Preparation Failed in Kubernetes

  #314 opened 7 months ago by Syulin7
  0

• fsdp not supported for Llama model

  #276 opened 9 months ago by jeffnvidia
  0

• llama2 7b SFT met error

  #271 opened 9 months ago by inspurasc
  0

• nemo container link in the README is not accessible any more

  #247 opened 10 months ago by Fizzbb
  1

• Failed to execute a multirun with different configurations in K8S

  #238 opened 10 months ago by Syulin7
  2

• k8s llama2 Data_preparation error inquiry

  #236 opened 10 months ago by donggyulee1
  1

• k8s llama2 Data preparation error inquiry

  #235 opened 10 months ago by donggyulee1
  0

• Add support for all tasks in latest version of eval harness

  #212 opened a year ago by juletx
  0

• How do we enable samples per second metric in NeMo training?

  #55 opened 2 years ago by a-cavalcanti
  2

• OSError: sbatch: error: Batch job submission failed: Invalid generic resource (gres) specification

  #72 opened 2 years ago by starlitsky2010
  1

• GPT3 126m divergence

  #207 opened a year ago by ethanhe42
  0

• Docker Build Fails

  #184 opened a year ago by TaekyungHeo
  7

• Nvidia nemo blog and Apples to Oranges comparison of H200 vs A100

  #180 opened a year ago by Qubitium
  1

• Documentation Code Block Display Issue

  #178 opened a year ago by KTH1234
  0

• Nemo Ai

  #161 opened a year ago by GAN-007
  0

• Request of NeMo-Megatron-Launcher in other language for local inferencing

  #159 opened a year ago by bm777
  0

• Is nemo-rlhf code available?

  #53 opened 2 years ago by maxhgerlach
  1

• Does NeMo-Megatron-Launcher support training from bare metal environment

  #132 opened a year ago by zigzagcai
  1

• is CSP (Azure,Aws or etc..) must to work on NeMo framework for Generative AI work? Cant we scale/leverage it on -OnPremise database with 8 GPUs

  #112 opened a year ago by Shantan243
  0

• Add Transformer Engine and Apex commit to support matrix

  #87 opened 2 years ago by wdykas
  0

• AssertionError: FP8 amax reduction group is not initialized

  #81 opened 2 years ago by starlitsky2010
  1

• requests.exceptions.SSLError: HTTPSConnectionPool

  #75 opened 2 years ago by starlitsky2010
  0

• OSError: [Errno 30] Read-only file system: 'outputs'

  #74 opened 2 years ago by starlitsky2010
  0

• Which versions of Pyxis, Slurm and enroot for running NeMo-Megatron-Launcher on one Node with 8 * A100?

  #73 opened 2 years ago by starlitsky2010
  2

• Where to execute the python main.py when use slurm only on only one GPU Node ?

  #69 opened 2 years ago by starlitsky2010
  2

• pyxis: failed to import docker image

  #28 opened 2 years ago by szhengac
  1

• Access to the script for evaluating the inference performance (average latency vs. model size)

  #22 opened 2 years ago by songkq
  3

• Convert nemo-megatron-mt5-3B to a binary file of fastertransformer successfully, but tritonserver fails with undesired shape when loading models.

  #21 opened 2 years ago by songkq
  2

• Is there a contribution guide for this repo?

  #7 opened 2 years ago by kaiyux
  3

• Device Support

  #6 opened 2 years ago by MikeDean2367
  4