saif-4b / serving /TESTS.json
GyrTech's picture
Add CPU-verified decision-only serving kit and container usage
873a5a8 verified
Raw History Blame Contribute Delete
5.56 kB
{
"reference": [
{
"model": "saif-4b",
"answers": {
"satisfied": {
"type": "noul",
"probabilities": {
"true": 0.9977252035331371,
"false": 0.0022747964668629324
},
"noul": 0.9977252035331371,
"answer": true
}
},
"usage": {
"input_tokens": 154,
"output_tokens": 0
}
},
{
"model": "saif-4b",
"answers": {
"intent": {
"type": "choice",
"probabilities": {
"complaint": 0.0014628375324945344,
"praise": 0.9985371624675055
},
"choice": "praise",
"answer": "praise",
"confidence": 0.9985371624675055
}
},
"usage": {
"input_tokens": 141,
"output_tokens": 0
}
},
{
"model": "saif-4b",
"answers": {
"sentiment": {
"type": "score",
"probabilities": {
"0": 0.0029920151680415817,
"1": 0.06401839839575167,
"2": 0.9329895864362067
},
"score": 1.929997571268165,
"answer": 2,
"confidence": 0.9329895864362067,
"legend": {
"0": "negative",
"1": "neutral",
"2": "positive"
}
}
},
"usage": {
"input_tokens": 152,
"output_tokens": 0
}
}
],
"versions": {
"torch": "2.8.0+cu128",
"transformers": "5.17.0",
"safetensors": "0.8.0",
"fastapi": "0.142.2",
"uvicorn": "0.54.0",
"pydantic": "2.14.1",
"huggingface_hub": "1.33.0",
"accelerate": "1.14.0"
},
"gpu_tested": false,
"gpu_note": "CPU only; local Docker has runc only, no NVIDIA runtime.",
"routes": [
"/docs",
"/docs/oauth2-redirect",
"/health",
"/openapi.json",
"/redoc",
"/v1/systemone"
],
"no_generation_routes": true,
"64_question_schema_accepted": true,
"local": {
"passed": true,
"requests": [
{
"type": "noul",
"exact_match": true,
"probabilities_sum_one": true,
"latency_seconds": 2.8982914260122925
},
{
"type": "choice",
"exact_match": true,
"probabilities_sum_one": true,
"latency_seconds": 2.7930272528901696
},
{
"type": "score",
"exact_match": true,
"probabilities_sum_one": true,
"latency_seconds": 1.9185104710049927
}
],
"malformed_422_count": 9,
"generation_routes_404": true,
"python_version": "3.12.14"
},
"di_client_live_passed": true,
"64_question_live": {
"passed": true,
"latency_seconds": 91.86650505103171
},
"cli": {
"happy_path_passed": true,
"help_passed": true,
"missing_file_exit_code": 2
},
"lsp": "Unavailable: basedpyright not installed; installation previously declined",
"container": {
"passed": true,
"requests": [
{
"type": "noul",
"exact_match": true,
"probabilities_sum_one": true,
"latency_seconds": 3.9418553828727454
},
{
"type": "choice",
"exact_match": true,
"probabilities_sum_one": true,
"latency_seconds": 1.384989856975153
},
{
"type": "score",
"exact_match": true,
"probabilities_sum_one": true,
"latency_seconds": 1.695539910113439
}
],
"malformed_422_count": 9,
"generation_routes_404": true,
"user": "1000:1000",
"health": "healthy",
"python_version": "3.11 (official PyTorch base)",
"versions_match_local": true,
"initial_startup": "Default UID could not read owner-only 0600 weights. Re-run with non-root weight-owner UID:GID succeeded; documented --user flag. Permissions unchanged."
},
"docker_build": {
"passed": true,
"image_id": "sha256:576663ba2b8710c2906afdc7a6466a8f9de8003791225f124ff6ab1fa0aa3d76",
"default_user": "saif",
"base": "pytorch/pytorch:2.8.0-cuda12.8-cudnn9-runtime",
"weights_baked_in": false
},
"observations": [
"Existing release tokenizer emits incorrect-regex warning; loader/tokenizer unchanged to preserve verified model outputs.",
"HF startup-download path implemented, not exercised with a full redundant weight download."
],
"registry": {
"verified": true,
"repository_id": 577,
"tags": {
"0.1.0": {
"name": "0.1.0",
"path": "sultan/saif/saif-4b-serving:0.1.0",
"location": "registry.omaemirates.com/sultan/saif/saif-4b-serving:0.1.0",
"revision": "a8b7239c2d32287ecb02aa17c8659e63bcbd0a3273ed0620f1534f34f7baf81f",
"short_revision": "a8b7239c2",
"digest": "sha256:576663ba2b8710c2906afdc7a6466a8f9de8003791225f124ff6ab1fa0aa3d76",
"created_at": "2026-10-28T19:04:01.024+04:00",
"total_size": 4340402812
},
"latest": {
"name": "latest",
"path": "sultan/saif/saif-4b-serving:latest",
"location": "registry.omaemirates.com/sultan/saif/saif-4b-serving:latest",
"revision": "a8b7239c2d32287ecb02aa17c8659e63bcbd0a3273ed0620f1534f34f7baf81f",
"short_revision": "a8b7239c2",
"digest": "sha256:576663ba2b8710c2906afdc7a6466a8f9de8003791225f124ff6ab1fa0aa3d76",
"created_at": "2026-10-28T19:04:01.024+04:00",
"total_size": 4340402812
}
},
"push_method": "Registry v2 8 MiB PATCH chunks; Docker monolithic upload rejected HTTP 413",
"docker_pull_passed": true
}
}