happyme531

/

MiniCPM-V-2_6-rkllm

like 5

Hi I ran the local file you provided and the run_rknn.py example, but I get the error below.

E RKNN: [11:03:49.214] failed to read scale tensor: must read size(206925824), real read size(206922780)

I would like to know how to resolve this.

board : orange pi 5 plus
os : ubuntu 24.04
rknpu : 0.9.7

$ python3 run_rknn.py 
W rknn-toolkit-lite2 version: 2.2.0
Start loading vision encoder model (size: 942.29 MB)
Vision encoder loaded in 14.82 seconds (speed: 63.58 MB/s)
I RKNN: [11:01:25.690] RKNN Runtime Information, librknnrt version: 2.2.0 (c195366594@2024-09-14T12:18:56)
I RKNN: [11:01:25.693] RKNN Driver Information, version: 0.9.7
I RKNN: [11:01:25.709] RKNN Model Information, version: 6, toolkit version: 2.2.0(compiler version: 2.2.0 (c195366594@2024-09-14T12:24:14)), target: RKNPU v2, target platform: rk3588, framework name: ONNX, framework layout: NCHW, model inference type: dynamic_shape
(1, 448, 448, 3)
Vision encoder inference time: 4.85 seconds
  C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False

(1, 64, 3584)
Start loading language model (size: 8279.12 MB)
I rkllm: rkllm-runtime version: 1.1.0, rknpu driver version: 0.9.7, platform: RK3588

I RKNN: [11:01:52.189] RKNN Driver Information, version: 0.9.7
E RKNN: [11:03:49.214] failed to read scale tensor: must read size(206925824), real read size(206922780)
E RKNN: [11:03:49.216] load model file error!
rknn_init fail! ret=-1
Language model loaded in 140.95 seconds (speed: 58.74 MB/s)
make LLM model
Start inference...
Segmentation fault