Vladislav Bronzov 22e102ad98 Bug fix gguf qwen2moe (#33940) ...

* fix qwen2moe tensors mapping, add unit tests

* add expert tensor split logic, test refactoring

* small params refactoring

* add comment to tensor reshaping

2024-10-05 16:19:01 +02:00

..

aqlm_integration

Cache: use batch_size instead of max_batch_size (#32657)

2024-08-16 11:48:45 +01:00

autoawq

Enables CPU AWQ model with IPEX version. (#33460)

2024-10-04 16:25:10 +02:00

bnb

Enable BNB multi-backend support (#31098)

2024-09-24 03:40:56 -06:00

compressed_tensor

HFQuantizer implementation for compressed-tensors library (#31704)

2024-09-25 14:31:38 +02:00

eetq_integration

[FEAT]: EETQ quantizer support (#30262)

2024-04-22 20:38:58 +01:00

fbgemm_fp8

Fix FbgemmFp8Linear not preserving tensor shape (#33239)

2024-09-11 13:26:44 +02:00

ggml

Bug fix gguf qwen2moe (#33940)

2024-10-05 16:19:01 +02:00

gptq

🚨 Remove dataset with restrictive license (#31452)

2024-06-17 17:56:51 +01:00

hqq

Hqq serialization (#33141)

2024-09-30 14:47:18 +02:00

quanto_integration

[Quantization] Switch to optimum-quanto (#31732)

2024-10-02 15:14:34 +02:00

torchao_integration

Add TorchAOHfQuantizer (#32306)

2024-08-14 16:14:24 +02:00