trident-ec / sp3_spine_text_33M.json
farguney's picture
Upload sp3_spine_text_33M.json with huggingface_hub
268e5f3 verified
Raw
History Blame
6.36 kB
{
"screen": "SP-2-SPINE-LM",
"seed": 0,
"steps": 4000,
"batch": 32,
"block": 512,
"lrs": [
0.001
],
"text_mb": 96.0,
"d_spine": 640,
"layers_spine": 8,
"n_spine": 33137920,
"tf_mult": 1.0,
"d_transformer": 672,
"layers_transformer": 6,
"n_transformer": 33257952,
"device": "cuda",
"thresholds": {
"pass": 1.05,
"concern": 1.15
},
"corpora": {
"text": {
"spine_lr0.001": {
"best_bpb": 1.5229665040969849,
"history": [
{
"step": 0,
"train_bits": 8.129907461277208,
"eval_bpb": 5.5101518630981445,
"best": 5.5101518630981445,
"kBps": 11.0
},
{
"step": 250,
"train_bits": 2.1948707954659055,
"eval_bpb": 2.1676111221313477,
"best": 2.1676111221313477,
"kBps": 19.2
},
{
"step": 500,
"train_bits": 1.9935468724699963,
"eval_bpb": 1.9912030696868896,
"best": 1.9912030696868896,
"kBps": 19.1
},
{
"step": 750,
"train_bits": 1.8484287715852128,
"eval_bpb": 1.8803513050079346,
"best": 1.8803513050079346,
"kBps": 19.1
},
{
"step": 1000,
"train_bits": 1.8020306041048408,
"eval_bpb": 1.8245058059692383,
"best": 1.8245058059692383,
"kBps": 19.0
},
{
"step": 1250,
"train_bits": 1.7574234639505542,
"eval_bpb": 1.7559597492218018,
"best": 1.7559597492218018,
"kBps": 19.0
},
{
"step": 1500,
"train_bits": 1.692556251576676,
"eval_bpb": 1.7225568294525146,
"best": 1.7225568294525146,
"kBps": 19.0
},
{
"step": 1750,
"train_bits": 1.7061122680829843,
"eval_bpb": 1.6795181035995483,
"best": 1.6795181035995483,
"kBps": 19.0
},
{
"step": 2000,
"train_bits": 1.6381732735820456,
"eval_bpb": 1.6360663175582886,
"best": 1.6360663175582886,
"kBps": 19.0
},
{
"step": 2250,
"train_bits": 1.5964389316277712,
"eval_bpb": 1.6126888990402222,
"best": 1.6126888990402222,
"kBps": 19.0
},
{
"step": 2500,
"train_bits": 1.5226208821897382,
"eval_bpb": 1.5832231044769287,
"best": 1.5832231044769287,
"kBps": 19.0
},
{
"step": 2750,
"train_bits": 1.5009390293954639,
"eval_bpb": 1.5672656297683716,
"best": 1.5672656297683716,
"kBps": 19.0
},
{
"step": 3000,
"train_bits": 1.5241976191328486,
"eval_bpb": 1.5502691268920898,
"best": 1.5502691268920898,
"kBps": 19.0
},
{
"step": 3250,
"train_bits": 1.526238021302685,
"eval_bpb": 1.5373672246932983,
"best": 1.5373672246932983,
"kBps": 19.0
},
{
"step": 3500,
"train_bits": 1.4832688439152155,
"eval_bpb": 1.531510591506958,
"best": 1.531510591506958,
"kBps": 19.0
},
{
"step": 3750,
"train_bits": 1.4648457183894898,
"eval_bpb": 1.5232207775115967,
"best": 1.5232207775115967,
"kBps": 19.0
},
{
"step": 3999,
"train_bits": 1.51629054477678,
"eval_bpb": 1.5229665040969849,
"best": 1.5229665040969849,
"kBps": 19.0
}
],
"wall_s": 3449.4
},
"tf_lr0.001": {
"best_bpb": 2.060558557510376,
"history": [
{
"step": 0,
"train_bits": 8.209524421737033,
"eval_bpb": 7.171977996826172,
"best": 7.171977996826172,
"kBps": 26.1
},
{
"step": 250,
"train_bits": 3.5390039550066,
"eval_bpb": 3.59385085105896,
"best": 3.59385085105896,
"kBps": 42.1
},
{
"step": 500,
"train_bits": 3.5031717136646354,
"eval_bpb": 3.5455288887023926,
"best": 3.5455288887023926,
"kBps": 42.2
},
{
"step": 750,
"train_bits": 3.2674196156884063,
"eval_bpb": 3.293250322341919,
"best": 3.293250322341919,
"kBps": 42.3
},
{
"step": 1000,
"train_bits": 3.103596165747831,
"eval_bpb": 3.1184656620025635,
"best": 3.1184656620025635,
"kBps": 42.4
},
{
"step": 1250,
"train_bits": 2.900629808436182,
"eval_bpb": 2.898190975189209,
"best": 2.898190975189209,
"kBps": 42.4
},
{
"step": 1500,
"train_bits": 2.6630564141873156,
"eval_bpb": 2.622429609298706,
"best": 2.622429609298706,
"kBps": 42.5
},
{
"step": 1750,
"train_bits": 2.438966459766093,
"eval_bpb": 2.4253695011138916,
"best": 2.4253695011138916,
"kBps": 42.5
},
{
"step": 2000,
"train_bits": 2.2740543596455565,
"eval_bpb": 2.3043291568756104,
"best": 2.3043291568756104,
"kBps": 42.5
},
{
"step": 2250,
"train_bits": 2.150469002596749,
"eval_bpb": 2.223310708999634,
"best": 2.223310708999634,
"kBps": 42.5
},
{
"step": 2500,
"train_bits": 2.1707502286823965,
"eval_bpb": 2.1691296100616455,
"best": 2.1691296100616455,
"kBps": 42.6
},
{
"step": 2750,
"train_bits": 2.114052020241851,
"eval_bpb": 2.1417148113250732,
"best": 2.1417148113250732,
"kBps": 42.6
},
{
"step": 3000,
"train_bits": 2.1376967110304226,
"eval_bpb": 2.1043992042541504,
"best": 2.1043992042541504,
"kBps": 42.6
},
{
"step": 3250,
"train_bits": 2.0466229543007204,
"eval_bpb": 2.085282564163208,
"best": 2.085282564163208,
"kBps": 42.6
},
{
"step": 3500,
"train_bits": 2.0477953600316523,
"eval_bpb": 2.0702152252197266,
"best": 2.0702152252197266,
"kBps": 42.6
},
{
"step": 3750,
"train_bits": 2.022161001955249,
"eval_bpb": 2.0639238357543945,
"best": 2.0639238357543945,
"kBps": 42.6
},
{
"step": 3999,
"train_bits": 2.018942862592304,
"eval_bpb": 2.060558557510376,
"best": 2.060558557510376,
"kBps": 42.6
}
],
"wall_s": 1538.3
},
"spine_best": 1.5229665040969849,
"tf_best": 2.060558557510376,
"ratio": 0.739103724349904,
"verdict": "PASS"
}
},
"partial": false,
"verdict": "PASS"
}