whatdhack

Custom images

5

@dariopavllo , congratulations on your presentation at NIPS 2020. Interesting work. Have a few quick questions. 1. What exactly would be involved in using custom images to generate 3D meshes...

Wondering, why the transformer codes are missing from the v2 branch. Any plan of bringing them in in the future ?

Could NOT find MPI_CXX (missing: MPI_CXX_WORKS)

14

**Environment:** 1. Framework: TensorFlow 2. Framework version: 2.4 3. Horovod version: 0.20.0 4. MPI version: 5. CUDA version: N/A 6. NCCL version: N/A 7. Python version: 3.7 8. Spark /...

Enable AMP (Automatic Mixed Precision ) in Tensorflow Serving.

11

### Describe the problem the feature is intended to solve AMP accelerates inference significantly. ### Describe the solution A flag for enabling AMP ### Describe alternatives you've considered There is...

Wondering why was the next(batch_it) sess.run was not incorporated in the in the immediately following sess.run in run_training_step() ? https://github.com/google/ffn/blob/3ea523c5475bacc2108df0071a8004f71dfbab65/train.py#L682 https://github.com/google/ffn/blob/3ea523c5475bacc2108df0071a8004f71dfbab65/train.py#L684

## ❓ Questions and Help Hi, wondering if you could clarify the following. 1. Why the normalizing part is not ignorable here like in discrimination tasks. I guess MCMC/SGLD is...

Meta-Llama-3-70B-Instruct running out of memory on 8 A100-40GB

4

## Describe the bug Out of memory. Tried to allocate X.XX GiB ..... ### Minimal reproducible example I guess any A100 system with 8+ GPUs ```python python example_chat_completion.py ``` ###...

[Examples] Adding a tensordict and TorchRL version of the PyTorch example

10

## Description Adding a torchdict and torch rl version of the PyTorch example reinforcement_q_learning.py ## Motivation and Context Adds a simpler DQN example - [ ] I have raised an...

Replace tf.train.SessionRunHook by tf.compat.v1.train.SessionRunHook ?

1

**Environment:** 1. Framework: TensorFlow, 2. Framework version: 2.16 3. Horovod version: 0.28.1 4. MPI version: 5. CUDA version: 12.2 6. NCCL version: 7. Python version: 3.11.8 8. Spark / PySpark...

### Willingness to contribute Yes. I can contribute this feature independently. ### Proposal Summary LLMs and other models are trained by running over multiple nodes with multiple GPUs spanning days....