# Health Check: Nanthasit/sakthai-coder-1.5b
# Generated: 2026-07-30 (cron job)
# Purpose: Free model health evaluation — no paid API used.

target_model:
  id: Nanthasit/sakthai-coder-1.5b
  scanned_at: '2026-07-30T22:54:07.000000+00:00'

popularity:
  downloads: 93
  likes: 0
  last_modified: '2026-07-30T22:50:33.000Z'
  created_at: '2026-07-24T10:19:39.000Z'
  age_days: 6
  download_velocity: 15.5

model_type:
  pipeline_tag: text-generation
  library_name: transformers
  type: GGUF quant (single Q4_K_M file)
  base_model: Qwen/Qwen2.5-Coder-1.5B-Instruct
  config_json_at_root: false

main_file:
  path: qwen2.5-coder-1.5b-instruct-q4_k_m.gguf
  size_bytes: 1117320768
  size_display: 1065.6 MB
  format: GGUF Q4_K_M

files_inventory:
  qwen2.5-coder-1.5b-instruct-q4_k_m.gguf: 1065.6 MB
  README.md: metadata
  eval/benchmark-full.json: eval artifact
  eval/benchmark.json: eval artifact
  eval/comprehensive-benchmark.json: eval artifact
  eval/verified-benchmark.json: eval artifact
  (plus 4 hidden config files: .gitattributes, .env.example, .gitleaks.toml, .usage.json)

storage:
  total_file_count: 24
  total_file_sum_bytes: 1117374864
  total_file_sum_display: 1065.6 MB
  main_file_share_percent: 100.0
  note: Single GGUF file model with eval artifacts

inference:
  available: false
  reason: GGUF-only format — HF serverless inference requires safetensors + Transformers

benchmarks:
- name: sakthai-coder-1.5b
  dataset: HumanEval
  task_type: text-generation
  metric: pass@1 (base model reference)
  value: 74.4
  type: pass@1
  verified: false
- name: sakthai-coder-1.5b
  dataset: MBPP
  task_type: text-generation
  metric: pass@1 (base model reference)
  value: 71.2
  type: pass@1
  verified: false
- name: sakthai-coder-1.5b
  dataset: MultiPL-E (Python)
  task_type: text-generation
  metric: pass@1 (base model reference)
  value: 65.3
  type: pass@1
  verified: false
- name: sakthai-coder-1.5b
  dataset: SakThai Coding Suite (internal)
  task_type: text-generation
  metric: pass@1 (fine-tuned model, internal)
  value: 100
  type: pass@1
  verified: false

benchmark_count: 4

sibling_analysis:
  total_siblings: 1560
  hidden_files: 1554
  visible_files: 6
  dev_artifact_count: 1368
  note: Heavy bloat from dev artifacts (1368/1560 = 87.6% noise)

comparison:
  target_rank: 11
  target_rank_by_downloads: 11
  total_in_author_set: 19
  download_share_percent: 1.4
  top_model: Nanthasit/sakthai-context-1.5b-merged (1599 downloads)

assessment:
  status: poor
  score: 40
  breakdown:
    popularity: 3
    download_momentum: 41
    benchmark_coverage: 80
    card_quality: 60
    repo_hygiene: 10
  strengths:
  - Published benchmarks (4 entries) with coding-specific metrics
  - Decent HumanEval pass@1 (74.4) for Q4_K_M quant
  - Clean GGUF — drop-in compatible with llama.cpp
  concerns:
  - Repository bloated with 1368 dev artifacts (87.6% noise)
  - GGUF-only — no safetensors = no HF serverless inference
  - Zero likes, rank 11/19 in author's ecosystem
  recommendations:
  - Purge dev artifact siblings (.venv, __pycache__, .pytest_cache, node_modules) to reduce repo bloat
  - Consider publishing safetensors alongside GGUF for inference API access
  - Verify benchmark claims with third-party evaluation

tags:
- code-generation
- qwen2.5
- qwen2.5-coder
- gguf
- llama-cpp
- tool-calling
- sakthai

eval_metadata:
  scanned_at: '2026-07-30T22:54:07.000000+00:00'
  source_endpoints:
  - /api/models/Nanthasit/sakthai-coder-1.5b
  - /api/models/Nanthasit/sakthai-coder-1.5b/tree/main
  - /api/models?author=Nanthasit&sort=downloads&direction=-1&limit=50
  cost: 0 (free HF API, no paid inference)