ai-engineering-from-scratch/phases/19-capstone-projects/07-end-to-end-fine-tuning-pipeline/quiz.json at main · Amjad-AbuRmileh/ai-engineering-from-scratch · GitHub

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
{
  "lesson": "07-end-to-end-fine-tuning-pipeline",
  "title": "Capstone 07 — End-to-End Fine-Tuning Pipeline (Data to SFT to DPO to Serve)",
  "questions": [
    {
      "stage": "pre",
      "question": "What does the contamination check guard against during data preparation?",
      "options": [
        "Tokenizer drift between SFT and DPO stages",
        "GPU driver mismatches between training and serving",
        "GGUF version skew across llama.cpp builds",
        "Test-set leakage from public benchmarks such as MMLU-Pro and MT-Bench-v2 into training data"
      ],
      "correct": 3,
      "explanation": ""
    },
    {
      "stage": "pre",
      "question": "Why does the pipeline compose SFT then DPO (or GRPO) rather than DPO alone?",
      "options": [
        "Axolotl does not implement DPO",
        "SFT establishes domain behavior on labeled completions while DPO or GRPO aligns the model against preference pairs or verifiable rewards",
        "DPO requires a separate base model architecture",
        "DPO cannot run on quantized weights"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "What does EAGLE-3 contribute to the vLLM serving stage?",
      "options": [
        "A new prompt-caching layer",
        "An OPA policy for tool calls",
        "An automatic data dedup step",
        "Draft heads that predict N tokens ahead; the target verifies in one pass for 2-3x throughput"
      ],
      "correct": 3,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "Which metric reports how well a speculative-decoding draft aligns with the target model?",
      "options": [
        "Acceptance rate",
        "Coverage delta",
        "PSI",
        "Perplexity"
      ],
      "correct": 0,
      "explanation": ""
    },
    {
      "stage": "check",
      "question": "Which trio of quants does the pipeline ship for deployment flexibility?",
      "options": [
        "ONNX, CoreML, TFLite",
        "GPTQ-INT4-Marlin, AWQ-INT4, and GGUF-Q4_K_M",
        "INT8 only across three runtimes",
        "FP32, FP16, BF16"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "post",
      "question": "Which framework convention does the 2026 model card follow in this capstone?",
      "options": [
        "Datasheets for Datasets",
        "Model Openness Framework (MOF) 2026 template covering data, training, eval, safety, license, and reproducibility",
        "HuggingFace YAML front-matter only",
        "OpenAI's model card format"
      ],
      "correct": 1,
      "explanation": ""
    },
    {
      "stage": "post",
      "question": "Which HPA metric is used to autoscale the serving replicas?",
      "options": [
        "Queue-wait time",
        "CPU utilization",
        "GPU temperature",
        "Network egress bytes"
      ],
      "correct": 0,
      "explanation": ""
    }
  ]
}