-
Notifications
You must be signed in to change notification settings - Fork 68
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
VLM模型的视觉rknn加载不了,failed to malloc npu memory, size: 1427746880 #195
Comments
内存分配的问题,尝试在rkllm添加上这个参数 param.extend_param.base_domain_id = 1; |
请问这个domain代表什么呢?可以解释一下么
…----------回复的邮件信息----------
***@***.***> 于2025年2月27日(周四) 上午8:55写道:
内存分配的问题,尝试在rkllm添加上这个参数 param.extend_param.base_domain_id = 1;
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
waydong left a comment (airockchip/rknn-llm#195)
内存分配的问题,尝试在rkllm添加上这个参数 param.extend_param.base_domain_id = 1;
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
为什么视觉语言模型的catch需要很大?我这里不能跑的原因是catch要求太多导致直接segment fault?
…----------回复的邮件信息----------
***@***.***> 于2025年2月27日(周四) 上午8:55写道:
内存分配的问题,尝试在rkllm添加上这个参数 param.extend_param.base_domain_id = 1;
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
waydong left a comment (airockchip/rknn-llm#195)
内存分配的问题,尝试在rkllm添加上这个参数 param.extend_param.base_domain_id = 1;
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
内存按4G分段,一个domain就是4G。 |
加上这个参数设置还是有问题,我的模型并不大,只是3b。这个share我理解就是放模型和参数的。不知道这个used在推理时用于做什么?catch也很大,另外我分配了swap4g空间,好像也没有用到。free还是没有内存,能解释一下原因么
…----------回复的邮件信息----------
***@***.***> 于2025年3月5日(周三) 下午4:33写道:
请问这个domain代表什么呢?可以解释一下么
…
内存按4G分段,一个domain就是4G。
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
waydong left a comment (airockchip/rknn-llm#195)
请问这个domain代表什么呢?可以解释一下么
…
内存按4G分段,一个domain就是4G。
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
已经将Qwen2.5_VL_3B的视觉模型转化为rknn(大小为:1.34G),语言模型转化为rkllm(大小为:3.48G),但是运行qwen2_vl_2B的demo出现视觉模型加载不进去,报错如下:
I rkllm: rkllm-runtime version: 1.1.4, rknpu driver version: 0.9.8, platform: RK3588
rkllm init success
main: LLM Model loaded in 17759.32 ms
E RKNN: [01:13:18.648] failed to allocate handle, ret: -1, errno: 14, errstr: Bad address
E RKNN: [01:13:18.656] failed to malloc npu memory, size: 1427746880, flags: 0x2
E RKNN: [01:13:18.656] Import rknn model failed!
E RKNN: [01:13:18.656] rknn_init, load model failed!
rknn_init fail! ret=-6
init_imgenc fail! ret=-1 model_path=../qwen2_vl_3b_vision_rk3588.rknn
我的ddr内存为16G,不知道为什么不能加载?我转换qwen2_vl_7B模型也出现同样的错误,请问如何解决?这个npu memory与什么有关系呢?
The text was updated successfully, but these errors were encountered: