Search results

Cannot set version_counter for inference tensor

bugSomething isn't workingSomething isn't working

float8

quantize_quantize_ APIquantize_ API

Status: Open.

#3487 In pytorch/ao;

· Stonepia opened on Dec 15, 2025

Llama-3.2-1B MX_Tensor inference failed

bugSomething isn't workingSomething isn't working

triaged

Status: Open.

Bug

#3339 In pytorch/ao;

· xiaowangintel opened on Nov 14, 2025

FLUX FP8 Training Error

bugSomething isn't workingSomething isn't working

float8

Status: Open.

Bug

#3111 In pytorch/ao;

· faruknane opened on Oct 1, 2025

PT2 Qat eval

bugSomething isn't workingSomething isn't working

Status: Open.

#2980 In pytorch/ao;

· xiguadong opened on Sep 5, 2025

CI ROCM tests failing with "HW exception - GPU hang"

bugSomething isn't workingSomething isn't working

ci

module: rocm

Status: Open.

#2890 In pytorch/ao;

· danielvegamyhre opened on Aug 27, 2025

float8 training + HF transformer model + torch.compile + DDP does not work

bugSomething isn't workingSomething isn't working

Status: Open.

#2586 In pytorch/ao;

· vkuzo opened on Jul 23, 2025

Int8DynActInt4WeightQuantizer quantize() not propagating the scales_precision argument

bugSomething isn't workingSomething isn't working

Status: Open.

#2571 In pytorch/ao;

· HaroldBenoit opened on Jul 18, 2025

newer torchao breaks sglang?

bugSomething isn't workingSomething isn't working

high priority

triage review

Status: Open.

#2226 In pytorch/ao;

· jerryzh168 opened on May 19, 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

[Needs more investigation] <code>int8_weight_only</code> via <code>quantize_()</code> API on <code>torch.float16</code> models results in NaN values across multiple CPU architectures

bugSomething isn't workingSomething isn't working

quantize_quantize_ APIquantize_ API

Status: Open.

#1662 In pytorch/ao;

· vmpuri opened on Feb 4, 2025

The internal torchao tensor subclasses cause errors with torch.compile

bugSomething isn't workingSomething isn't working

Status: Open.

#1463 In pytorch/ao;

· JohnnyRacer opened on Dec 28, 2024

<code>unhashable type: non-nested SymInt</code>

autoquant

bugSomething isn't workingSomething isn't working

Status: Open.

#1381 In pytorch/ao;

· bhack opened on Dec 4, 2024