Commit
•
fd59336
1
Parent(s):
3bb4cdc
Training in progress, step 95000
Browse files- adapter_model.safetensors +1 -1
- last-checkpoint/adapter_config.json +3 -3
- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/tokenizer_config.json +1 -0
- last-checkpoint/trainer_state.json +3 -423
- last-checkpoint/training_args.bin +2 -2
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5544997664
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c8c2bc861c786b2bf9d8ddb8858babedad8fc42c6e26fb00fe13b35096c6de7
|
| 3 |
size 5544997664
|
last-checkpoint/adapter_config.json
CHANGED
|
@@ -20,13 +20,13 @@
|
|
| 20 |
"rank_pattern": {},
|
| 21 |
"revision": null,
|
| 22 |
"target_modules": [
|
|
|
|
| 23 |
"up_proj",
|
|
|
|
| 24 |
"k_proj",
|
| 25 |
"gate_proj",
|
| 26 |
-
"down_proj",
|
| 27 |
-
"o_proj",
|
| 28 |
"v_proj",
|
| 29 |
-
"
|
| 30 |
],
|
| 31 |
"task_type": "CAUSAL_LM",
|
| 32 |
"use_dora": false,
|
|
|
|
| 20 |
"rank_pattern": {},
|
| 21 |
"revision": null,
|
| 22 |
"target_modules": [
|
| 23 |
+
"down_proj",
|
| 24 |
"up_proj",
|
| 25 |
+
"q_proj",
|
| 26 |
"k_proj",
|
| 27 |
"gate_proj",
|
|
|
|
|
|
|
| 28 |
"v_proj",
|
| 29 |
+
"o_proj"
|
| 30 |
],
|
| 31 |
"task_type": "CAUSAL_LM",
|
| 32 |
"use_dora": false,
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5544997664
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:588a15f7cea8d0a814b43f40bdcbefaf553e3f1ea3fe8f93ab138197ad9ca78c
|
| 3 |
size 5544997664
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 674093138
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ebee6fad5e5c1bb9d7e488846379921805f8c0c20003d8a91f2e25aed77a83a
|
| 3 |
size 674093138
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6d8f8f9743ec8d3f95f6d09874e8c8e1665b1753c549b2fad6b80c9e2a59f8a6
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:def38347a8476c414f9a59b485e01231d01480373c9bf5d7882acb65a1218490
|
| 3 |
size 1064
|
last-checkpoint/tokenizer_config.json
CHANGED
|
@@ -2072,6 +2072,7 @@
|
|
| 2072 |
"bos_token": "<|im_start|>",
|
| 2073 |
"chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
|
| 2074 |
"clean_up_tokenization_spaces": true,
|
|
|
|
| 2075 |
"eos_token": "<|im_end|>",
|
| 2076 |
"max_length": 4096,
|
| 2077 |
"model_input_names": [
|
|
|
|
| 2072 |
"bos_token": "<|im_start|>",
|
| 2073 |
"chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
|
| 2074 |
"clean_up_tokenization_spaces": true,
|
| 2075 |
+
"device_map": "auto",
|
| 2076 |
"eos_token": "<|im_end|>",
|
| 2077 |
"max_length": 4096,
|
| 2078 |
"model_input_names": [
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -6230,426 +6230,6 @@
|
|
| 6230 |
"learning_rate": 1.9991374234676826e-05,
|
| 6231 |
"loss": 1.5551,
|
| 6232 |
"step": 88900
|
| 6233 |
-
},
|
| 6234 |
-
{
|
| 6235 |
-
"epoch": 0.06618773518956465,
|
| 6236 |
-
"grad_norm": 0.6733551621437073,
|
| 6237 |
-
"learning_rate": 1.999135481992186e-05,
|
| 6238 |
-
"loss": 1.4334,
|
| 6239 |
-
"step": 89000
|
| 6240 |
-
},
|
| 6241 |
-
{
|
| 6242 |
-
"epoch": 0.06626210343135068,
|
| 6243 |
-
"grad_norm": 0.8035016059875488,
|
| 6244 |
-
"learning_rate": 1.999133538335166e-05,
|
| 6245 |
-
"loss": 1.4872,
|
| 6246 |
-
"step": 89100
|
| 6247 |
-
},
|
| 6248 |
-
{
|
| 6249 |
-
"epoch": 0.06633647167313671,
|
| 6250 |
-
"grad_norm": 0.4339046776294708,
|
| 6251 |
-
"learning_rate": 1.9991315924966277e-05,
|
| 6252 |
-
"loss": 1.4869,
|
| 6253 |
-
"step": 89200
|
| 6254 |
-
},
|
| 6255 |
-
{
|
| 6256 |
-
"epoch": 0.06641083991492273,
|
| 6257 |
-
"grad_norm": 0.6680594086647034,
|
| 6258 |
-
"learning_rate": 1.9991296444765747e-05,
|
| 6259 |
-
"loss": 1.5103,
|
| 6260 |
-
"step": 89300
|
| 6261 |
-
},
|
| 6262 |
-
{
|
| 6263 |
-
"epoch": 0.06648520815670876,
|
| 6264 |
-
"grad_norm": 0.697487473487854,
|
| 6265 |
-
"learning_rate": 1.9991276942750117e-05,
|
| 6266 |
-
"loss": 1.4239,
|
| 6267 |
-
"step": 89400
|
| 6268 |
-
},
|
| 6269 |
-
{
|
| 6270 |
-
"epoch": 0.06655957639849479,
|
| 6271 |
-
"grad_norm": 0.587734043598175,
|
| 6272 |
-
"learning_rate": 1.9991257418919424e-05,
|
| 6273 |
-
"loss": 1.5856,
|
| 6274 |
-
"step": 89500
|
| 6275 |
-
},
|
| 6276 |
-
{
|
| 6277 |
-
"epoch": 0.06663394464028081,
|
| 6278 |
-
"grad_norm": 0.8574571013450623,
|
| 6279 |
-
"learning_rate": 1.999123787327372e-05,
|
| 6280 |
-
"loss": 1.4818,
|
| 6281 |
-
"step": 89600
|
| 6282 |
-
},
|
| 6283 |
-
{
|
| 6284 |
-
"epoch": 0.06670831288206684,
|
| 6285 |
-
"grad_norm": 1.0861676931381226,
|
| 6286 |
-
"learning_rate": 1.9991218305813035e-05,
|
| 6287 |
-
"loss": 1.4883,
|
| 6288 |
-
"step": 89700
|
| 6289 |
-
},
|
| 6290 |
-
{
|
| 6291 |
-
"epoch": 0.06678268112385287,
|
| 6292 |
-
"grad_norm": 1.0139306783676147,
|
| 6293 |
-
"learning_rate": 1.9991198716537422e-05,
|
| 6294 |
-
"loss": 1.5099,
|
| 6295 |
-
"step": 89800
|
| 6296 |
-
},
|
| 6297 |
-
{
|
| 6298 |
-
"epoch": 0.0668570493656389,
|
| 6299 |
-
"grad_norm": 0.6741511225700378,
|
| 6300 |
-
"learning_rate": 1.999117910544692e-05,
|
| 6301 |
-
"loss": 1.4746,
|
| 6302 |
-
"step": 89900
|
| 6303 |
-
},
|
| 6304 |
-
{
|
| 6305 |
-
"epoch": 0.06693141760742492,
|
| 6306 |
-
"grad_norm": 0.9702801704406738,
|
| 6307 |
-
"learning_rate": 1.999115947254157e-05,
|
| 6308 |
-
"loss": 1.5166,
|
| 6309 |
-
"step": 90000
|
| 6310 |
-
},
|
| 6311 |
-
{
|
| 6312 |
-
"epoch": 0.06700578584921095,
|
| 6313 |
-
"grad_norm": 0.7757803797721863,
|
| 6314 |
-
"learning_rate": 1.9991139817821416e-05,
|
| 6315 |
-
"loss": 1.5031,
|
| 6316 |
-
"step": 90100
|
| 6317 |
-
},
|
| 6318 |
-
{
|
| 6319 |
-
"epoch": 0.06708015409099698,
|
| 6320 |
-
"grad_norm": 0.7200698256492615,
|
| 6321 |
-
"learning_rate": 1.9991120141286502e-05,
|
| 6322 |
-
"loss": 1.5834,
|
| 6323 |
-
"step": 90200
|
| 6324 |
-
},
|
| 6325 |
-
{
|
| 6326 |
-
"epoch": 0.067154522332783,
|
| 6327 |
-
"grad_norm": 0.7415780425071716,
|
| 6328 |
-
"learning_rate": 1.999110044293687e-05,
|
| 6329 |
-
"loss": 1.5689,
|
| 6330 |
-
"step": 90300
|
| 6331 |
-
},
|
| 6332 |
-
{
|
| 6333 |
-
"epoch": 0.06722889057456903,
|
| 6334 |
-
"grad_norm": 0.5777677297592163,
|
| 6335 |
-
"learning_rate": 1.9991080722772564e-05,
|
| 6336 |
-
"loss": 1.5139,
|
| 6337 |
-
"step": 90400
|
| 6338 |
-
},
|
| 6339 |
-
{
|
| 6340 |
-
"epoch": 0.06730325881635506,
|
| 6341 |
-
"grad_norm": 0.6991866827011108,
|
| 6342 |
-
"learning_rate": 1.999106098079363e-05,
|
| 6343 |
-
"loss": 1.5073,
|
| 6344 |
-
"step": 90500
|
| 6345 |
-
},
|
| 6346 |
-
{
|
| 6347 |
-
"epoch": 0.06737762705814108,
|
| 6348 |
-
"grad_norm": 0.6112390160560608,
|
| 6349 |
-
"learning_rate": 1.9991041217000105e-05,
|
| 6350 |
-
"loss": 1.4773,
|
| 6351 |
-
"step": 90600
|
| 6352 |
-
},
|
| 6353 |
-
{
|
| 6354 |
-
"epoch": 0.06745199529992713,
|
| 6355 |
-
"grad_norm": 0.8287676572799683,
|
| 6356 |
-
"learning_rate": 1.9991021431392033e-05,
|
| 6357 |
-
"loss": 1.5425,
|
| 6358 |
-
"step": 90700
|
| 6359 |
-
},
|
| 6360 |
-
{
|
| 6361 |
-
"epoch": 0.06752636354171315,
|
| 6362 |
-
"grad_norm": 0.8582881689071655,
|
| 6363 |
-
"learning_rate": 1.999100162396946e-05,
|
| 6364 |
-
"loss": 1.5581,
|
| 6365 |
-
"step": 90800
|
| 6366 |
-
},
|
| 6367 |
-
{
|
| 6368 |
-
"epoch": 0.06760073178349918,
|
| 6369 |
-
"grad_norm": 0.5585276484489441,
|
| 6370 |
-
"learning_rate": 1.999098179473243e-05,
|
| 6371 |
-
"loss": 1.5015,
|
| 6372 |
-
"step": 90900
|
| 6373 |
-
},
|
| 6374 |
-
{
|
| 6375 |
-
"epoch": 0.0676751000252852,
|
| 6376 |
-
"grad_norm": 0.4237435460090637,
|
| 6377 |
-
"learning_rate": 1.9990961943680984e-05,
|
| 6378 |
-
"loss": 1.523,
|
| 6379 |
-
"step": 91000
|
| 6380 |
-
},
|
| 6381 |
-
{
|
| 6382 |
-
"epoch": 0.06774946826707123,
|
| 6383 |
-
"grad_norm": 0.5455594658851624,
|
| 6384 |
-
"learning_rate": 1.999094207081517e-05,
|
| 6385 |
-
"loss": 1.5448,
|
| 6386 |
-
"step": 91100
|
| 6387 |
-
},
|
| 6388 |
-
{
|
| 6389 |
-
"epoch": 0.06782383650885726,
|
| 6390 |
-
"grad_norm": 0.48855817317962646,
|
| 6391 |
-
"learning_rate": 1.999092217613502e-05,
|
| 6392 |
-
"loss": 1.4535,
|
| 6393 |
-
"step": 91200
|
| 6394 |
-
},
|
| 6395 |
-
{
|
| 6396 |
-
"epoch": 0.06789820475064329,
|
| 6397 |
-
"grad_norm": 0.5199916958808899,
|
| 6398 |
-
"learning_rate": 1.999090225964059e-05,
|
| 6399 |
-
"loss": 1.4921,
|
| 6400 |
-
"step": 91300
|
| 6401 |
-
},
|
| 6402 |
-
{
|
| 6403 |
-
"epoch": 0.06797257299242931,
|
| 6404 |
-
"grad_norm": 0.5790271162986755,
|
| 6405 |
-
"learning_rate": 1.9990882321331916e-05,
|
| 6406 |
-
"loss": 1.5773,
|
| 6407 |
-
"step": 91400
|
| 6408 |
-
},
|
| 6409 |
-
{
|
| 6410 |
-
"epoch": 0.06804694123421534,
|
| 6411 |
-
"grad_norm": 0.5524342656135559,
|
| 6412 |
-
"learning_rate": 1.9990862361209043e-05,
|
| 6413 |
-
"loss": 1.4619,
|
| 6414 |
-
"step": 91500
|
| 6415 |
-
},
|
| 6416 |
-
{
|
| 6417 |
-
"epoch": 0.06812130947600137,
|
| 6418 |
-
"grad_norm": 0.7153291702270508,
|
| 6419 |
-
"learning_rate": 1.999084237927202e-05,
|
| 6420 |
-
"loss": 1.6042,
|
| 6421 |
-
"step": 91600
|
| 6422 |
-
},
|
| 6423 |
-
{
|
| 6424 |
-
"epoch": 0.0681956777177874,
|
| 6425 |
-
"grad_norm": 0.957635223865509,
|
| 6426 |
-
"learning_rate": 1.9990822375520882e-05,
|
| 6427 |
-
"loss": 1.538,
|
| 6428 |
-
"step": 91700
|
| 6429 |
-
},
|
| 6430 |
-
{
|
| 6431 |
-
"epoch": 0.06827004595957342,
|
| 6432 |
-
"grad_norm": 0.38240477442741394,
|
| 6433 |
-
"learning_rate": 1.9990802349955678e-05,
|
| 6434 |
-
"loss": 1.5937,
|
| 6435 |
-
"step": 91800
|
| 6436 |
-
},
|
| 6437 |
-
{
|
| 6438 |
-
"epoch": 0.06834441420135945,
|
| 6439 |
-
"grad_norm": 0.8961233496665955,
|
| 6440 |
-
"learning_rate": 1.999078230257645e-05,
|
| 6441 |
-
"loss": 1.5119,
|
| 6442 |
-
"step": 91900
|
| 6443 |
-
},
|
| 6444 |
-
{
|
| 6445 |
-
"epoch": 0.06841878244314548,
|
| 6446 |
-
"grad_norm": 0.47433900833129883,
|
| 6447 |
-
"learning_rate": 1.999076223338324e-05,
|
| 6448 |
-
"loss": 1.5449,
|
| 6449 |
-
"step": 92000
|
| 6450 |
-
},
|
| 6451 |
-
{
|
| 6452 |
-
"epoch": 0.0684931506849315,
|
| 6453 |
-
"grad_norm": 0.8222399353981018,
|
| 6454 |
-
"learning_rate": 1.9990742142376098e-05,
|
| 6455 |
-
"loss": 1.5334,
|
| 6456 |
-
"step": 92100
|
| 6457 |
-
},
|
| 6458 |
-
{
|
| 6459 |
-
"epoch": 0.06856751892671753,
|
| 6460 |
-
"grad_norm": 0.464373916387558,
|
| 6461 |
-
"learning_rate": 1.999072202955506e-05,
|
| 6462 |
-
"loss": 1.5003,
|
| 6463 |
-
"step": 92200
|
| 6464 |
-
},
|
| 6465 |
-
{
|
| 6466 |
-
"epoch": 0.06864188716850356,
|
| 6467 |
-
"grad_norm": 0.8799763321876526,
|
| 6468 |
-
"learning_rate": 1.9990701894920176e-05,
|
| 6469 |
-
"loss": 1.581,
|
| 6470 |
-
"step": 92300
|
| 6471 |
-
},
|
| 6472 |
-
{
|
| 6473 |
-
"epoch": 0.06871625541028958,
|
| 6474 |
-
"grad_norm": 0.9567086100578308,
|
| 6475 |
-
"learning_rate": 1.999068173847149e-05,
|
| 6476 |
-
"loss": 1.4373,
|
| 6477 |
-
"step": 92400
|
| 6478 |
-
},
|
| 6479 |
-
{
|
| 6480 |
-
"epoch": 0.06879062365207561,
|
| 6481 |
-
"grad_norm": 0.440479576587677,
|
| 6482 |
-
"learning_rate": 1.999066156020904e-05,
|
| 6483 |
-
"loss": 1.5571,
|
| 6484 |
-
"step": 92500
|
| 6485 |
-
},
|
| 6486 |
-
{
|
| 6487 |
-
"epoch": 0.06886499189386165,
|
| 6488 |
-
"grad_norm": 0.7486180663108826,
|
| 6489 |
-
"learning_rate": 1.9990641360132876e-05,
|
| 6490 |
-
"loss": 1.4437,
|
| 6491 |
-
"step": 92600
|
| 6492 |
-
},
|
| 6493 |
-
{
|
| 6494 |
-
"epoch": 0.06893936013564768,
|
| 6495 |
-
"grad_norm": 0.7576742172241211,
|
| 6496 |
-
"learning_rate": 1.9990621138243037e-05,
|
| 6497 |
-
"loss": 1.5306,
|
| 6498 |
-
"step": 92700
|
| 6499 |
-
},
|
| 6500 |
-
{
|
| 6501 |
-
"epoch": 0.0690137283774337,
|
| 6502 |
-
"grad_norm": 0.6755186915397644,
|
| 6503 |
-
"learning_rate": 1.9990600894539574e-05,
|
| 6504 |
-
"loss": 1.5769,
|
| 6505 |
-
"step": 92800
|
| 6506 |
-
},
|
| 6507 |
-
{
|
| 6508 |
-
"epoch": 0.06908809661921973,
|
| 6509 |
-
"grad_norm": 0.6093853712081909,
|
| 6510 |
-
"learning_rate": 1.9990580629022526e-05,
|
| 6511 |
-
"loss": 1.5777,
|
| 6512 |
-
"step": 92900
|
| 6513 |
-
},
|
| 6514 |
-
{
|
| 6515 |
-
"epoch": 0.06916246486100576,
|
| 6516 |
-
"grad_norm": 0.5788242220878601,
|
| 6517 |
-
"learning_rate": 1.9990560341691938e-05,
|
| 6518 |
-
"loss": 1.494,
|
| 6519 |
-
"step": 93000
|
| 6520 |
-
},
|
| 6521 |
-
{
|
| 6522 |
-
"epoch": 0.06923683310279179,
|
| 6523 |
-
"grad_norm": 0.828676700592041,
|
| 6524 |
-
"learning_rate": 1.9990540032547855e-05,
|
| 6525 |
-
"loss": 1.5651,
|
| 6526 |
-
"step": 93100
|
| 6527 |
-
},
|
| 6528 |
-
{
|
| 6529 |
-
"epoch": 0.06931120134457781,
|
| 6530 |
-
"grad_norm": 0.5612863302230835,
|
| 6531 |
-
"learning_rate": 1.9990519701590322e-05,
|
| 6532 |
-
"loss": 1.5584,
|
| 6533 |
-
"step": 93200
|
| 6534 |
-
},
|
| 6535 |
-
{
|
| 6536 |
-
"epoch": 0.06938556958636384,
|
| 6537 |
-
"grad_norm": 0.965107262134552,
|
| 6538 |
-
"learning_rate": 1.999049934881938e-05,
|
| 6539 |
-
"loss": 1.497,
|
| 6540 |
-
"step": 93300
|
| 6541 |
-
},
|
| 6542 |
-
{
|
| 6543 |
-
"epoch": 0.06945993782814987,
|
| 6544 |
-
"grad_norm": 0.46939852833747864,
|
| 6545 |
-
"learning_rate": 1.9990478974235078e-05,
|
| 6546 |
-
"loss": 1.5716,
|
| 6547 |
-
"step": 93400
|
| 6548 |
-
},
|
| 6549 |
-
{
|
| 6550 |
-
"epoch": 0.0695343060699359,
|
| 6551 |
-
"grad_norm": 0.4986964464187622,
|
| 6552 |
-
"learning_rate": 1.999045857783746e-05,
|
| 6553 |
-
"loss": 1.5762,
|
| 6554 |
-
"step": 93500
|
| 6555 |
-
},
|
| 6556 |
-
{
|
| 6557 |
-
"epoch": 0.06960867431172192,
|
| 6558 |
-
"grad_norm": 0.4267128109931946,
|
| 6559 |
-
"learning_rate": 1.9990438159626566e-05,
|
| 6560 |
-
"loss": 1.5101,
|
| 6561 |
-
"step": 93600
|
| 6562 |
-
},
|
| 6563 |
-
{
|
| 6564 |
-
"epoch": 0.06968304255350795,
|
| 6565 |
-
"grad_norm": 0.411811888217926,
|
| 6566 |
-
"learning_rate": 1.9990417719602445e-05,
|
| 6567 |
-
"loss": 1.5623,
|
| 6568 |
-
"step": 93700
|
| 6569 |
-
},
|
| 6570 |
-
{
|
| 6571 |
-
"epoch": 0.06975741079529398,
|
| 6572 |
-
"grad_norm": 0.8761053681373596,
|
| 6573 |
-
"learning_rate": 1.999039725776514e-05,
|
| 6574 |
-
"loss": 1.4294,
|
| 6575 |
-
"step": 93800
|
| 6576 |
-
},
|
| 6577 |
-
{
|
| 6578 |
-
"epoch": 0.06983177903708,
|
| 6579 |
-
"grad_norm": 0.9531000852584839,
|
| 6580 |
-
"learning_rate": 1.99903767741147e-05,
|
| 6581 |
-
"loss": 1.4925,
|
| 6582 |
-
"step": 93900
|
| 6583 |
-
},
|
| 6584 |
-
{
|
| 6585 |
-
"epoch": 0.06990614727886603,
|
| 6586 |
-
"grad_norm": 0.516830325126648,
|
| 6587 |
-
"learning_rate": 1.999035626865116e-05,
|
| 6588 |
-
"loss": 1.5802,
|
| 6589 |
-
"step": 94000
|
| 6590 |
-
},
|
| 6591 |
-
{
|
| 6592 |
-
"epoch": 0.06998051552065206,
|
| 6593 |
-
"grad_norm": 0.47061294317245483,
|
| 6594 |
-
"learning_rate": 1.9990335741374572e-05,
|
| 6595 |
-
"loss": 1.5668,
|
| 6596 |
-
"step": 94100
|
| 6597 |
-
},
|
| 6598 |
-
{
|
| 6599 |
-
"epoch": 0.07005488376243808,
|
| 6600 |
-
"grad_norm": 0.7790777683258057,
|
| 6601 |
-
"learning_rate": 1.9990315192284978e-05,
|
| 6602 |
-
"loss": 1.5568,
|
| 6603 |
-
"step": 94200
|
| 6604 |
-
},
|
| 6605 |
-
{
|
| 6606 |
-
"epoch": 0.07012925200422411,
|
| 6607 |
-
"grad_norm": 0.75156170129776,
|
| 6608 |
-
"learning_rate": 1.9990294621382426e-05,
|
| 6609 |
-
"loss": 1.5217,
|
| 6610 |
-
"step": 94300
|
| 6611 |
-
},
|
| 6612 |
-
{
|
| 6613 |
-
"epoch": 0.07020362024601014,
|
| 6614 |
-
"grad_norm": 1.195028305053711,
|
| 6615 |
-
"learning_rate": 1.999027402866696e-05,
|
| 6616 |
-
"loss": 1.5662,
|
| 6617 |
-
"step": 94400
|
| 6618 |
-
},
|
| 6619 |
-
{
|
| 6620 |
-
"epoch": 0.07027798848779618,
|
| 6621 |
-
"grad_norm": 0.6215851306915283,
|
| 6622 |
-
"learning_rate": 1.999025341413862e-05,
|
| 6623 |
-
"loss": 1.5208,
|
| 6624 |
-
"step": 94500
|
| 6625 |
-
},
|
| 6626 |
-
{
|
| 6627 |
-
"epoch": 0.0703523567295822,
|
| 6628 |
-
"grad_norm": 0.509843647480011,
|
| 6629 |
-
"learning_rate": 1.9990232777797458e-05,
|
| 6630 |
-
"loss": 1.489,
|
| 6631 |
-
"step": 94600
|
| 6632 |
-
},
|
| 6633 |
-
{
|
| 6634 |
-
"epoch": 0.07042672497136823,
|
| 6635 |
-
"grad_norm": 1.2951029539108276,
|
| 6636 |
-
"learning_rate": 1.9990212119643516e-05,
|
| 6637 |
-
"loss": 1.4729,
|
| 6638 |
-
"step": 94700
|
| 6639 |
-
},
|
| 6640 |
-
{
|
| 6641 |
-
"epoch": 0.07050109321315426,
|
| 6642 |
-
"grad_norm": 0.5028135776519775,
|
| 6643 |
-
"learning_rate": 1.9990191439676838e-05,
|
| 6644 |
-
"loss": 1.5579,
|
| 6645 |
-
"step": 94800
|
| 6646 |
-
},
|
| 6647 |
-
{
|
| 6648 |
-
"epoch": 0.07057546145494029,
|
| 6649 |
-
"grad_norm": 0.7202877998352051,
|
| 6650 |
-
"learning_rate": 1.9990170737897473e-05,
|
| 6651 |
-
"loss": 1.5282,
|
| 6652 |
-
"step": 94900
|
| 6653 |
}
|
| 6654 |
],
|
| 6655 |
"logging_steps": 100,
|
|
@@ -6669,7 +6249,7 @@
|
|
| 6669 |
"attributes": {}
|
| 6670 |
}
|
| 6671 |
},
|
| 6672 |
-
"total_flos": 1.
|
| 6673 |
"train_batch_size": 1,
|
| 6674 |
"trial_name": null,
|
| 6675 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.06611336694777863,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 88900,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 6230 |
"learning_rate": 1.9991374234676826e-05,
|
| 6231 |
"loss": 1.5551,
|
| 6232 |
"step": 88900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6233 |
}
|
| 6234 |
],
|
| 6235 |
"logging_steps": 100,
|
|
|
|
| 6249 |
"attributes": {}
|
| 6250 |
}
|
| 6251 |
},
|
| 6252 |
+
"total_flos": 1.211678461812007e+18,
|
| 6253 |
"train_batch_size": 1,
|
| 6254 |
"trial_name": null,
|
| 6255 |
"trial_params": null
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8bb89e5e4b20648cd50836d8df065bde229d29cb5c6085310a18725c84aab824
|
| 3 |
+
size 5496
|