Search results

int8 quantization with FSDP for inference error

distributed

quantize

triaged

Status: Open.

#2127 In pytorch/ao;

· Andy0422 opened on Apr 25, 2025

Dynamo error with large mesh + AdamWFp8 + bf16 stochastic rounding

bugSomething isn't workingSomething isn't working

distributed

high priority

optimizer

triage review

triaged

Status: Open.

#2074 In pytorch/ao;

· cassanof opened on Apr 18, 2025

[Bug] FSDP2 FP8 compatibility problem with nn.Linear layers (GPU count > out_features)

distributed

float8

triaged

Status: Open.

#1938 In pytorch/ao;

· HIT-cwh opened on Mar 24, 2025

[QAT] Low-bit FSDP all-gather for QAT

distributed

good first issueGood for newcomersGood for newcomers

qat

Status: Open.

#1224 In pytorch/ao;

· gau-nernst opened on Nov 5, 2024

[NF4][FSDP2]: enable multi-gpu CI

ci

distributed

Status: Open.

#202 In pytorch/ao;

· weifengpy opened on May 2, 2024