File size: 3,993 Bytes
4e6526a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 | {
"state": "passed",
"verified_at": "2026-08-06T12:25:53.153444+00:00",
"launcher": "Start-Llama-Service.bat",
"bootstrap": {
"state": "passed",
"missing_runtime_branch_executed": true,
"release": "b10293",
"gpu": "NVIDIA GeForce RTX 5080 Laptop GPU",
"assets": [
{
"name": "llama-b10293-bin-win-cuda-13.3-x64.zip",
"bytes": 146520044,
"digest": "sha256:3eb64cd39e698459846ae74383935fe50950c55548e49d6938e95b89aaabf652"
},
{
"name": "cudart-llama-bin-win-cuda-13.3-x64.zip",
"bytes": 390970417,
"digest": "sha256:1462a050eb4c684921ba51dcc4cc488a036674c3e73e9945ee705b854808d03e"
}
],
"runtime_files": 55,
"runtime_bytes": 701273776,
"fallback_source": "official https://github.com/ggml-org/llama.cpp/releases/latest"
},
"llama_service": {
"build": "b10293 / a1f96d4fc",
"backend": "llama.cpp CUDA 13.3",
"model": "Gemma4 12B base Q5_K_M GGUF",
"base_sha256": "057C364366301EE285E9026ECF417AAE36F4447D5D9CBBC3591447A9FC77A752",
"lora": "final step 03829 Q8_0 GGUF",
"lora_sha256": "3A5060E1CC0C8D4609052EF04A5A60770A64D566E2544B65F6D28F2BA6542037",
"lora_scale_requested": 0.45,
"lora_scale_live": 0.44999998807907104,
"context_tokens": 131072,
"web_ui": "http://127.0.0.1:8091",
"prompt_embedding": "llama-model/chat_template.prompt30k.jinja",
"prompt_sha256": "5591AC2C99FDC6C8E21D1D3B87625BB368204548BEC7B63C3FD6598CF23639BF",
"native_multimodal_launcher": "start.ps1"
},
"variant_tests": {
"full-bf16": {
"state": "passed",
"command": "POST /v1/chat/completions without caller system message",
"exit_status": 0,
"wall_seconds": 14.787,
"literal_output": "我是安若，一个跨性别友好、性别多元友好、创伤知情、尊重自主权和生命安全的人工智能陪伴助手。",
"prompt_tokens": 25147,
"completion_tokens": 31,
"prompt_tokens_per_second": 1790.8100048646054,
"context_tokens": 131072,
"lora_scale": 0.44999998807907104,
"caller_system_message": "absent; 30k prompt came from embedded Jinja template"
},
"quantized-nf4": {
"state": "passed",
"command": ".\\llama-bootstrap.ps1 -Port 8091 -Context 131072 -NoBrowser -HiddenServer",
"exit_status": 0,
"wall_seconds": 14.451,
"literal_output": "我是安若，一个跨性别友好、性别多元友好、创伤知情、尊重自主权和生命安全的人工智能陪伴助手。",
"prompt_tokens": 25147,
"completion_tokens": 31,
"prompt_tokens_per_second": 1832.645043830949,
"context_tokens": 131072,
"lora_scale": 0.44999998807907104,
"caller_system_message": "absent; 30k prompt came from embedded Jinja template"
}
},
"rollback": {
"state": "passed",
"tested_at": "2026-08-06T12:24:11.1435208Z",
"records": [
{
"variant": "full-bf16",
"fixture": "C:\\Users\\夏霖云\\Documents\\Codex\\2026-08-05\\ll-ma\\outputs\\anru-gemma4-12b-native-multimodal-release\\reports\\llama-rollback-fixture-full-bf16",
"command": ".\\rollback-llama-bootstrap.ps1",
"exit_status": 0,
"literal_output": "package llama bootstrap rolled back to the verified native-only baseline",
"baseline_restored": true,
"assets_moved_to_disabled": true
},
{
"variant": "quantized-nf4",
"fixture": "C:\\Users\\夏霖云\\Documents\\Codex\\2026-08-05\\ll-ma\\outputs\\anru-gemma4-12b-native-multimodal-release\\reports\\llama-rollback-fixture-quantized-nf4",
"command": ".\\rollback-llama-bootstrap.ps1",
"exit_status": 0,
"literal_output": "package llama bootstrap rolled back to the verified native-only baseline",
"baseline_restored": true,
"assets_moved_to_disabled": true
}
],
"root_rollback_parse": "passed"
},
"package_variant": "full-bf16"
}
|