Search results

[Feature Req] Can you add *args and **kwargs to improve extensibility ?

triaged

Status: Open.

#2496 In pytorch/ao;

· Musa-Sina-Ertugrul opened on Jul 6, 2025

[MoE training] crash with FSDP if shared_expert uses float8 in torchtitan llama4

triaged

Status: Open.

#2453 In pytorch/ao;

· danielvegamyhre opened on Jun 27, 2025

TorchAO needs to update its build system

build

triaged

Status: Open.

#2222 In pytorch/ao;

· drisspg opened on May 19, 2025

[QAT] Linear layer's weight quantization granularity can only be per_group

qat

triaged

Status: Open.

#2189 In pytorch/ao;

· Cypher-Bruce opened on May 9, 2025

python setup.develop is taking a lot of time

build

triaged

Status: Open.

#2150 In pytorch/ao;

· anijain2305 opened on Apr 29, 2025

[roadmap/tracker] Low precision MoE training

tracker

triaged

Status: Open.

#2147 In pytorch/ao;

· danielvegamyhre opened on Apr 29, 2025

AO/GemLite tensors produce incorrect outputs in vLLM

integrationIssues related to integrations with other libraries, like huggingface, vllm, sglang, gemlite etc.Issues related to integrations with other libraries, like huggingface, vllm, sglang, gemlite etc.

quantize_quantize_ APIquantize_ API

triaged

Status: Open.

#2141 In pytorch/ao;

· mobicham opened on Apr 28, 2025

AssertionError During Quantization of <code>torch.empty_like()</code>, <code>torch.ones_like</code>, and <code>torch.randn_like</code>

pt2e_quantpt2 export quantization (prepare_pt2e, convert_pt2e)pt2 export quantization (prepare_pt2e, convert_pt2e)

triaged

Status: Open.

#2146 In pytorch/ao;

· defaultd661 opened on Apr 28, 2025

QAT model drops accuracy after converting with torch.ao.quantization.convert

qat

triaged

Status: Open.

#2138 In pytorch/ao;

· tranngocduvnvp opened on Apr 28, 2025

int8 quantization with FSDP for inference error

distributed

quantize_quantize_ APIquantize_ API

triaged

Status: Open.

#2127 In pytorch/ao;

· Andy0422 opened on Apr 25, 2025

Question about dtype check in marlin_qqq validation for w4a8 functionality

quantize_quantize_ APIquantize_ API

triaged

Status: Open.

#2115 In pytorch/ao;

· xxw11 opened on Apr 23, 2025

[PT2E] observers do not handle inputs with different shapes correctly

pt2e_quantpt2 export quantization (prepare_pt2e, convert_pt2e)pt2 export quantization (prepare_pt2e, convert_pt2e)

triaged

Status: Open.

#2112 In pytorch/ao;

· Xia-Weiwen opened on Apr 23, 2025