Views
No views yet
cminst/realistic-bpe5-general-mix-specialist-1b-20b.
Source run directory name: R5FW20B-spec-general_mix_bpe32-1B-20B-tillicum.cminst/realistic-bpe5-general-mix-20b.model.pt: final model weightsconfig.json: training/model configuration saved by train_v3.pyresults.json: final evaluation metrics and logsstep_*/: intermediate model-only checkpoints (model.pt, config.json, results.json, and string_atlas.json when present)tokenizers/*.json: tokenizer JSON filesgeneral_mix_bpe32general_mix_bpe32: BPB 0.8535 at step 190831{
2 "anchor": "general_mix_bpe32",
3 "anchor_weight": 1.0,
4 "completed_steps": 19083,
5 "ctx_len": 2048,
6 "d_model": 2048,
7 "data_format": "packed_tokens",
8 "input_mode": "full",
9 "lr_decay_steps": 19083,
10 "model_size": "1B",
11 "n_heads": 32,
12 "n_layers": 20,
13 "output_mode": "full",
14 "run_name": "R5FW20B-spec-general_mix_bpe32-1B-20B-tillicum",
15 "tokens_processed": 20000205312,
16 "vocab_sizes": {
17 "general_mix_bpe32": 32000
18 }
19}