Token Classification
Transformers
Safetensors
Ancient Greek (to 1453)
char_bert_meter
ancient-greek
classical-philology
character-level
macronization
metrical-scansion
custom_code
Instructions to use Ericu950/CharDiff-grc-meter with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Ericu950/CharDiff-grc-meter with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="Ericu950/CharDiff-grc-meter", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Ericu950/CharDiff-grc-meter", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "cfg": { | |
| "name": "meter_joint_docclean", | |
| "seed": 0, | |
| "ckpt": "/nobackup/proj/disk/dionysus/shared/GreekCharBERT/runs/gcb_doc_clean/best.pt", | |
| "out_dir": "/nobackup/proj/disk/dionysus/shared/GreekCharBERT/runs/meter_joint_docclean", | |
| "encoded": "/nobackup/proj/disk/dionysus/shared/GreekCharBERT/meter/encoded", | |
| "attn": "sdpa", | |
| "T": 2048, | |
| "micro_batch": 16, | |
| "eval_micro": 16, | |
| "lr_enc": 0.0001, | |
| "lr_head": 0.0008, | |
| "wd": 0.01, | |
| "clip": 1.0, | |
| "warmup_frac": 0.1, | |
| "epochs": 25, | |
| "patience": 6, | |
| "head_dropout": 0.33, | |
| "use_cap": true, | |
| "scalar_mix": true, | |
| "w_mac": 1.0, | |
| "w_scan": 1.0, | |
| "select": "mean", | |
| "prose_per_epoch": 150000, | |
| "scan_passes": 2, | |
| "scan_max_verses": 8, | |
| "mac_class_w": "auto", | |
| "scan_class_w": "auto", | |
| "log_every": 50 | |
| }, | |
| "mcfg": { | |
| "head_dropout": 0.33, | |
| "w_mac": 1.0, | |
| "w_scan": 1.0, | |
| "use_cap": true, | |
| "scalar_mix": true, | |
| "mac_class_w": [ | |
| 3.935332098208277, | |
| 0.5727731680236648 | |
| ], | |
| "scan_class_w": [ | |
| 0.4482191825355474, | |
| 1.2170555073099973, | |
| 1.2004924282537328, | |
| 8.748756624191806 | |
| ] | |
| }, | |
| "pretrain_cfg": { | |
| "name": "gcb_doc_clean", | |
| "seed": 0, | |
| "d_model": 1024, | |
| "depth": 32, | |
| "char_window": 256, | |
| "attn": "flex", | |
| "compile": true, | |
| "qk_norm": true, | |
| "seq_len": 8192, | |
| "micro_batch": 4, | |
| "grad_accum": 2, | |
| "lr": 0.0003, | |
| "wd": 0.1, | |
| "clip": 1.0, | |
| "warmup_frac": 0.04, | |
| "decay_frac": 0.4, | |
| "tier_weights": [ | |
| 1.0, | |
| 1.0, | |
| 0.3 | |
| ], | |
| "lam": 0.1, | |
| "total_steps": 120000, | |
| "hard_max_steps": 1000000, | |
| "num_workers": 16, | |
| "log_every": 20, | |
| "ckpt_every": 2000, | |
| "eval_every_ckpt": true, | |
| "eval_n": 256, | |
| "out_dir": "/nobackup/proj/disk/dionysus/shared/GreekCharBERT/runs/gcb_doc_clean", | |
| "anneal_phases": [ | |
| [ | |
| 0.5, | |
| [ | |
| 1.0, | |
| 1.0, | |
| 0.3 | |
| ] | |
| ], | |
| [ | |
| 0.9, | |
| [ | |
| 1.0, | |
| 1.0, | |
| 0.0 | |
| ] | |
| ], | |
| [ | |
| 1.0, | |
| [ | |
| 1.0, | |
| 0.0, | |
| 0.0 | |
| ] | |
| ] | |
| ] | |
| }, | |
| "epoch": 12, | |
| "dev": { | |
| "mac_bal": 0.8927, | |
| "mac_acc": 0.9195, | |
| "mac_n": 2498, | |
| "scan_bal": 0.9811, | |
| "scan_acc": 0.9787, | |
| "end_f1": 0.9791, | |
| "scan_n": 68004 | |
| }, | |
| "T": 2048, | |
| "_source_checkpoint": "/nobackup/proj/disk/dionysus/shared/GreekCharBERT/runs/meter_joint_docclean/best.pt" | |
| } |