Instructions to use md-zain/resume-parser with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use md-zain/resume-parser with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/Qwen2.5-3B-Instruct-bnb-4bit") model = PeftModel.from_pretrained(base_model, "md-zain/resume-parser") - Notebooks
- Google Colab
- Kaggle
Download evaluation_summary.json from md-zain/resume-parser: direct link, hf CLI and curl.
- Browser
- Download file 3.71 kB
-
https://huggingface.co/md-zain/resume-parser/resolve/main/evaluation_summary.json
- Command line
-
hf download hf://md-zain/resume-parser/evaluation_summary.json
-
curl -L -o evaluation_summary.json https://huggingface.co/md-zain/resume-parser/resolve/main/evaluation_summary.json
3.71 kB
| { | |
| "model_repo": "md-zain/resume-parser", | |
| "evaluated_model_commit": "86649ee4bc8b06987bbd98462b009892183b2e2a", | |
| "source_checkpoint_commit": "4128d0e4627ee5d71f88c4e9f14efc4f147f64c2", | |
| "evaluation_status": "completed_400_resume_sample", | |
| "dataset_repo": "md-zain/resume-parser-dataset", | |
| "split": "test", | |
| "source_split_sha256": "f01b3831d19f8528683c629521c7d618a439aedc050accaf98028d6bd2471f9a", | |
| "sampling": "uniform random sample without replacement; seed 42; original split row order", | |
| "generation": { | |
| "max_new_tokens": 2048, | |
| "context_length": 32768, | |
| "do_sample": false, | |
| "load_in_4bit": true, | |
| "compute_dtype": "float16", | |
| "prompt_style": "compact", | |
| "stop": "EOS or first complete JSON checked every 16 tokens" | |
| }, | |
| "prompt_sha256": "ca308193030079391c13bfc8772649004c4bf999d0861f6ab47af004264423c6", | |
| "schema_sha256": "45bd363dd4c97a256a48163084e26b2c5f389c377ff08751a89f35226f5cbb61", | |
| "counts": { | |
| "items_evaluated": 400, | |
| "valid_targets": 400, | |
| "json_valid_count": 357, | |
| "schema_valid_count": 356, | |
| "scored_count": 356, | |
| "source_rows": 1463, | |
| "expected_rows": 400, | |
| "completed_rows": 400, | |
| "sample_seed": 42, | |
| "retryable_error_count": 0, | |
| "output_limit_count": 43 | |
| }, | |
| "scores": { | |
| "mean_programmatic_score": 0.908859, | |
| "mean_semantic_score": 0.806636, | |
| "mean_overall_score": 0.857748, | |
| "mean_identity_score": 0.978901, | |
| "overall_score_including_invalid_predictions": 0.763395, | |
| "json_valid_rate": 0.8925, | |
| "schema_valid_rate": 0.89 | |
| }, | |
| "status_counts": { | |
| "SCORED": 356, | |
| "INVALID_JSON": 43, | |
| "INVALID_PREDICTION_SCHEMA": 1 | |
| }, | |
| "scoring": "Step 04 Evaluator: deterministic field comparisons plus TypeSafe AI semantic scoring", | |
| "score_interpretation": { | |
| "mean_overall_score": "Mean among the 356 fully scored resumes.", | |
| "overall_score_including_invalid_predictions": "Mean over all 400 valid targets, with failed outputs scored as zero." | |
| }, | |
| "limitations": [ | |
| "This is a 400-resume sample of 1463 reserved test resumes, not full-split or universal accuracy.", | |
| "All 43 invalid JSON outputs hit the declared 2048-token generation ceiling.", | |
| "Historical pre-evaluation runs each used 5 resumes with different samples/prompts/settings; comparison is descriptive only.", | |
| "The Quantized label follows the historical filename; exact precision is not embedded in that report." | |
| ], | |
| "historical_comparison": { | |
| "runs": [ | |
| { | |
| "name": "Pre-fine-tuning (Quantized)", | |
| "evaluated": 5, | |
| "scored": 3, | |
| "json_valid_count": 4, | |
| "schema_valid_count": 3, | |
| "overall": 0.719074, | |
| "programmatic": 0.774263, | |
| "semantic": 0.663885, | |
| "identity": 1.0, | |
| "overall_all": 0.431444, | |
| "json_valid": 0.8, | |
| "schema_valid": 0.6 | |
| }, | |
| { | |
| "name": "Post-fine-tuning", | |
| "evaluated": 400, | |
| "scored": 356, | |
| "json_valid_count": 357, | |
| "schema_valid_count": 356, | |
| "overall": 0.857748, | |
| "programmatic": 0.908859, | |
| "semantic": 0.806636, | |
| "identity": 0.978901, | |
| "overall_all": 0.763395, | |
| "json_valid": 0.8925, | |
| "schema_valid": 0.89 | |
| } | |
| ], | |
| "difference_vs_quantized_report_percentage_points": { | |
| "overall_all": 33.195100000000004, | |
| "overall": 13.867399999999996, | |
| "programmatic": 13.459599999999995, | |
| "semantic": 14.275100000000007, | |
| "identity": -2.109899999999998, | |
| "json_valid": 9.249999999999991, | |
| "schema_valid": 29.000000000000004 | |
| }, | |
| "matched_evaluation_conditions": false | |
| }, | |
| "wandb_run": "https://wandb.ai/zainmemon010-personal/resume-parser/runs/do783w9g" | |
| } | |