Search results

v0.6.1 does not compile SM89 (Ada Lovelace) kernels unlike v0.5.3

Status: Open.

#1956 In kvcache-ai/ktransformers;

· y8nqd4jtfn-cell opened on Apr 26, 2026

对Qwen3-235B-A22B 官方FP16版本支持有问题，模型会胡说八道。

Status: Open.

#1949 In kvcache-ai/ktransformers;

· 3139725181 opened on Apr 25, 2026

[New Model]: Support Deepseek V4 Series

enhancementNew feature or requestNew feature or request

Status: Open.

#1946 In kvcache-ai/ktransformers;

· tunglinwood opened on Apr 24, 2026

GPTQ_INT4 method crashes with RuntimeError on MoE weight loading (w2 tensor dimension mismatch)

Status: Open.

#1943 In kvcache-ai/ktransformers;

· lyncker opened on Apr 23, 2026

kt-numa-nodes 配置单numa运行，cpu，gpu均按照numa切分，ttft耗时极高

Status: Open.

#1938 In kvcache-ai/ktransformers;

· poryfly opened on Apr 21, 2026

Pipeline Parallel for ktransformers for multi low-vram GPUs

enhancementNew feature or requestNew feature or request

Status: Open.

#1933 In kvcache-ai/ktransformers;

· NukePeace opened on Apr 16, 2026

FP8 crashes on sm_86 (RTX 3090) even with --kt-num-gpu-experts 0

Status: Open.

#1930 In kvcache-ai/ktransformers;

· ormandj opened on Apr 13, 2026

是否有计划支持gemma4 26B模型

enhancementNew feature or requestNew feature or request

Status: Open.

#1924 In kvcache-ai/ktransformers;

· shown1985 opened on Apr 9, 2026

BUG: main(1dd0a7889901097555262d3ff3521018a2e3ece8): Fatal Python error: Segmentation fault

Status: Open.

#1922 In kvcache-ai/ktransformers;

· Honghe opened on Apr 9, 2026

KTransformers Roadmap (2026 Q2)

enhancementNew feature or requestNew feature or request

Status: Open.

#1921 In kvcache-ai/ktransformers;

· ErvinXie opened on Apr 9, 2026

BF16 CPUInfer on Hygon C86 + BW150 produces layer-2 NaNs and then crashes with HSA_STATUS_ERROR_MEMORY_APERTURE_VIOLATION

Status: Open.

#1917 In kvcache-ai/ktransformers;

· skyguan92 opened on Apr 7, 2026

在L20*8上部署了GLM5-FP8，感觉速度有点慢，想看看能不能再优化一下

Status: Open.

#1915 In kvcache-ai/ktransformers;

· applezzj opened on Apr 3, 2026