LLM on local device, can I Skip the file upload?

[scsonic]

https://github.com/quic/ai-hub-apps/tree/main/tutorials/llm_on_genie

When I execute the command:
python -m qai_hub_models.models.llama_v3_8b_chat_quantized.export --device "Snapdragon 8 Elite QRD" --skip-inferencing --skip-profiling --output-dir genie_bundle

I do not want to upload the files because it wastes several hours.
Can I just convert the model, save it locally on Ubuntu, and then use adb push to transfer it to Android?

It seems like I must complete the upload process to generate the binary model files
(e.g., llama_v3_8b_chat_quantized_part_X_of_X.bin).
Is that correct?

[scsonic]

INFO:backoff:Backing off _try_upload_asset(...) for 2.6s (requests.exceptions.SSLError: HTTPSConnectionPool(host='tetrahub-qprod-userdata.s3.amazonaws.com', port=443): Max retries exceeded with url: /models/xxxxxxxxxxxx.aimet.zip?uploadId=xxxxxxxxxxxxxx-&partNumber=2&AWSAccessKeyId=xxxxxxx&Signature=xxxxxxxx&x-amz-security-token=xxxxxxxx(Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:2426)'))))
1%|█▎ | 9.56M/1.00G [00:27<46:40, 380kB/s]

if you get a error, network error or anything else, all your upload progress was gone and you have no idea to skip prev finished job
the compile and quantized must be process online?

[1826133674]

I have also encountered a similar error. Have you resolved it? How did you solve it?

INFO:backoff:Backing off _try_upload_asset(...) for 2.6s (requests.exceptions.SSLError: HTTPSConnectionPool(host='tetrahub-qprod-userdata.s3.amazonaws.com', port=443): Max retries exceeded with url: /models/xxxxxxxxxxxx.aimet.zip?uploadId=xxxxxxxxxxxxxx-&partNumber=2&AWSAccessKeyId=xxxxxxx&Signature=xxxxxxxx&x-amz-security-token=xxxxxxxx(Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:2426)')))) 1%|█▎ | 9.56M/1.00G [00:27<46:40, 380kB/s]

if you get a error, network error or anything else, all your upload progress was gone and you have no idea to skip prev finished job the compile and quantized must be process online?

[1826133674]

INFO:backoff:Backing off _try_upload_asset(...) for 2.6s (requests.exceptions.SSLError: HTTPSConnectionPool(host='tetrahub-qprod-userdata.s3.amazonaws.com', port=443): Max retries exceeded with url: /models/xxxxxxxxxxxx.aimet.zip?uploadId=xxxxxxxxxxxxxx-&partNumber=2&AWSAccessKeyId=xxxxxxx&Signature=xxxxxxxx&x-amz-security-token=xxxxxxxx(Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:2426)')))) 1%|█▎ | 9.56M/1.00G [00:27<46:40, 380kB/s]

if you get a error, network error or anything else, all your upload progress was gone and you have no idea to skip prev finished job the compile and quantized must be process online?

Additionally, I have found that this error recurs repeatedly. And it always happens during the upload process you mentioned.I tried three times, and the situation was the same each time. Is this actually a bug?

[pkhara31]

I also face the same error. Snapshot below. I also tried it 4-5 times but it fails everytime.

[scsonic]

finally,
I used the company's network to the upload files, more than 10GB
and I received three .bin files in return.

I have uploaded them to this Google Drive for reference.
You must use QNN 2.28.2.241116.
If you download these .bin files, you can directly build ChatApp on macOS.

https://drive.google.com/drive/folders/19FHlmAi9Ucr0UbPN5v46XcGLPpyMkgiP?usp=sharing

[1826133674]

finally, I used the company's network to the upload files, more than 10GB and I received three .bin files in return.

I have uploaded them to this Google Drive for reference. You must use QNN 2.28.2.241116. If you download these .bin files, you can directly build ChatApp on macOS.

https://drive.google.com/drive/folders/19FHlmAi9Ucr0UbPN5v46XcGLPpyMkgiP?usp=sharing

Thanks for your sharing! I tried it myself and made it to the step of downloading the model at the end. However, I failed to download the third part of the model. Have you encountered this issue?

The error is as follows:
Waiting for compile job (jgjvoryeg) completion. Type Ctrl+C to stop waiting at any time.
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=4, connect=5, read=4, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/jgjvoryeg/
✅ SUCCESS
llama_v3_2_3b_chat_quantized_part_1_of_3.bin: 100%|█████████████████████████████████| 752M/752M [02:47<00:00, 4.69MB/s]
Downloaded model to genie_bundle\llama_v3_2_3b_chat_quantized_part_1_of_3.bin
llama_v3_2_3b_chat_quantized_part_2_of_3.bin: 100%|█████████████████████████████████| 688M/688M [02:29<00:00, 4.83MB/s]
Downloaded model to genie_bundle\llama_v3_2_3b_chat_quantized_part_2_of_3.bin
Waiting for link job (j5we8lzm5) completion. Type Ctrl+C to stop waiting at any time.
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=4, connect=5, read=4, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/j5we8lzm5/
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=4, connect=5, read=4, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/j5we8lzm5/
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=3, connect=5, read=3, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/j5we8lzm5/
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=4, connect=5, read=4, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/j5we8lzm5/
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=3, connect=5, read=3, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/j5we8lzm5/
WARNING:urllib3.connectionpool:Retrying (LogRetry(total=2, connect=5, read=2, redirect=None, status=None)) after connection broken by 'ConnectionResetError(10054, '远程主机强迫关闭了一个现有的连接。', None, 10054, None)': /api/v1/jobs/j5we8lzm5/
❌ FAILED Internal linking error