train_wsc_1745950303

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the wsc dataset. It achieves the following results on the evaluation set:

• Loss: 0.5956
• Num Input Tokens Seen: 14002704

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 123
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• training_steps: 40000

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.9367	1.6024	200	0.6859	70144
0.7729	3.2008	400	0.6358	140304
0.6178	4.8032	600	0.6251	210240
0.7354	6.4016	800	0.6166	279952
0.8205	8.0	1000	0.6166	350224
0.9947	9.6024	1200	0.6063	420256
0.8109	11.2008	1400	0.6140	490496
0.6329	12.8032	1600	0.6054	560224
0.6829	14.4016	1800	0.6053	630560
0.6086	16.0	2000	0.6093	699648
0.485	17.6024	2200	0.6015	769232
0.9604	19.2008	2400	0.6084	839344
0.6535	20.8032	2600	0.6110	909744
0.6409	22.4016	2800	0.6115	979312
0.7109	24.0	3000	0.6066	1049184
0.7251	25.6024	3200	0.6034	1119552
0.6356	27.2008	3400	0.6066	1189008
0.8557	28.8032	3600	0.6137	1259168
0.759	30.4016	3800	0.6130	1329056
0.9193	32.0	4000	0.6128	1399280
0.7954	33.6024	4200	0.6092	1469920
0.7279	35.2008	4400	0.6051	1539184
0.941	36.8032	4600	0.6034	1609648
0.9295	38.4016	4800	0.6008	1679792
0.7476	40.0	5000	0.6098	1749008
0.8862	41.6024	5200	0.6106	1818832
0.7252	43.2008	5400	0.6087	1889136
0.501	44.8032	5600	0.6182	1959008
0.4602	46.4016	5800	0.6046	2028320
0.7075	48.0	6000	0.6129	2098928
0.7795	49.6024	6200	0.6080	2168688
0.6954	51.2008	6400	0.6075	2238752
0.905	52.8032	6600	0.6000	2308816
0.8237	54.4016	6800	0.6067	2379328
0.6337	56.0	7000	0.6052	2448704
0.8776	57.6024	7200	0.6037	2519008
0.7921	59.2008	7400	0.6066	2588608
0.8712	60.8032	7600	0.6045	2659072
0.6104	62.4016	7800	0.6041	2728480
0.9738	64.0	8000	0.6079	2798720
0.6123	65.6024	8200	0.6013	2868672
0.5486	67.2008	8400	0.6026	2939312
0.4234	68.8032	8600	0.6083	3009568
0.706	70.4016	8800	0.6032	3079584
0.5217	72.0	9000	0.6046	3149680
0.4153	73.6024	9200	0.6172	3219680
0.4354	75.2008	9400	0.6041	3289472
0.6993	76.8032	9600	0.5956	3359520
0.7275	78.4016	9800	0.6037	3429568
0.5396	80.0	10000	0.6079	3499648
0.7598	81.6024	10200	0.6038	3569504
0.7379	83.2008	10400	0.6109	3639920
0.9387	84.8032	10600	0.6056	3709520
0.7098	86.4016	10800	0.5983	3779456
0.6795	88.0	11000	0.6039	3849744
0.7353	89.6024	11200	0.6032	3919984
0.6685	91.2008	11400	0.6080	3989872
0.7216	92.8032	11600	0.6073	4059568
0.8336	94.4016	11800	0.6013	4129664
0.548	96.0	12000	0.6024	4199936
0.9363	97.6024	12200	0.5981	4269952
0.6282	99.2008	12400	0.6110	4339040
0.7682	100.8032	12600	0.6031	4409680
0.9204	102.4016	12800	0.6103	4479120
0.6169	104.0	13000	0.6119	4548896
0.7145	105.6024	13200	0.6044	4619216
0.7454	107.2008	13400	0.6099	4689424
0.7114	108.8032	13600	0.6078	4759232
0.7552	110.4016	13800	0.6081	4829120
0.5361	112.0	14000	0.6138	4899024
0.6323	113.6024	14200	0.5998	4968944
0.7257	115.2008	14400	0.6055	5039152
0.5306	116.8032	14600	0.6010	5109312
0.8061	118.4016	14800	0.6115	5179296
0.7583	120.0	15000	0.6079	5249504
0.818	121.6024	15200	0.6016	5319424
0.909	123.2008	15400	0.6039	5389488
0.9621	124.8032	15600	0.6032	5459776
0.3719	126.4016	15800	0.6107	5529760
0.8277	128.0	16000	0.6074	5599968
0.5884	129.6024	16200	0.6056	5671056
0.6286	131.2008	16400	0.6104	5740000
0.6262	132.8032	16600	0.6098	5810288
0.6929	134.4016	16800	0.6065	5880176
0.6835	136.0	17000	0.6080	5950048
0.7025	137.6024	17200	0.6135	6020016
0.8546	139.2008	17400	0.6162	6090672
0.5158	140.8032	17600	0.6072	6160288
0.7597	142.4016	17800	0.6078	6230656
0.8127	144.0	18000	0.6005	6299968
0.669	145.6024	18200	0.6080	6370512
0.7968	147.2008	18400	0.6064	6440784
0.5663	148.8032	18600	0.6056	6510560
0.6785	150.4016	18800	0.6010	6579872
0.8551	152.0	19000	0.6024	6650112
0.7856	153.6024	19200	0.5996	6720368
0.5416	155.2008	19400	0.6072	6790512
0.7651	156.8032	19600	0.6056	6860880
0.6543	158.4016	19800	0.6175	6930576
0.5508	160.0	20000	0.6053	7000640
0.6528	161.6024	20200	0.6023	7070272
0.6598	163.2008	20400	0.5996	7140336
0.5761	164.8032	20600	0.6078	7210816
0.653	166.4016	20800	0.6016	7281392
0.8061	168.0	21000	0.6057	7350960
0.7621	169.6024	21200	0.6053	7421312
0.6579	171.2008	21400	0.6047	7491200
0.5762	172.8032	21600	0.6003	7560976
0.9284	174.4016	21800	0.6020	7631024
0.6199	176.0	22000	0.6054	7700784
0.7859	177.6024	22200	0.6110	7770752
0.3245	179.2008	22400	0.6039	7840832
0.7359	180.8032	22600	0.6061	7911072
0.7983	182.4016	22800	0.6075	7981312
0.6592	184.0	23000	0.6066	8050976
0.6686	185.6024	23200	0.6060	8121312
0.5448	187.2008	23400	0.6047	8191520
0.5868	188.8032	23600	0.6013	8261456
0.7454	190.4016	23800	0.6131	8331664
1.137	192.0	24000	0.6159	8401328
0.5008	193.6024	24200	0.6039	8471232
0.8048	195.2008	24400	0.6079	8540976
0.6897	196.8032	24600	0.6059	8611296
0.5966	198.4016	24800	0.6075	8681264
0.434	200.0	25000	0.6160	8751280
0.4255	201.6024	25200	0.6050	8822192
0.5553	203.2008	25400	0.6063	8891648
0.6894	204.8032	25600	0.6118	8961760
0.5924	206.4016	25800	0.6104	9031568
0.4732	208.0	26000	0.6030	9101088
0.7517	209.6024	26200	0.6052	9171168
0.3247	211.2008	26400	0.6049	9240752
0.5487	212.8032	26600	0.6017	9310960
0.7838	214.4016	26800	0.6027	9380560
1.0043	216.0	27000	0.6075	9450912
0.4924	217.6024	27200	0.6063	9520832
0.5188	219.2008	27400	0.6075	9590800
0.826	220.8032	27600	0.6111	9661456
0.9029	222.4016	27800	0.6089	9731376
0.5354	224.0	28000	0.6084	9801040
0.6485	225.6024	28200	0.6080	9870784
0.8221	227.2008	28400	0.6132	9941408
0.7324	228.8032	28600	0.6031	10011264
0.7633	230.4016	28800	0.6112	10080704
0.9061	232.0	29000	0.6090	10150880
0.855	233.6024	29200	0.6018	10221616
0.9609	235.2008	29400	0.6006	10291664
0.7309	236.8032	29600	0.6120	10361728
0.7132	238.4016	29800	0.6046	10431088
0.5857	240.0	30000	0.6083	10501088
0.6568	241.6024	30200	0.6097	10571488
0.8502	243.2008	30400	0.6069	10640848
0.7067	244.8032	30600	0.6096	10711136
0.5737	246.4016	30800	0.6039	10781136
0.411	248.0	31000	0.5998	10851312
0.3786	249.6024	31200	0.6112	10921664
0.8119	251.2008	31400	0.6060	10991936
0.7882	252.8032	31600	0.6012	11061680
0.7779	254.4016	31800	0.6105	11131872
0.5879	256.0	32000	0.6011	11201520
0.4562	257.6024	32200	0.6092	11271952
0.8154	259.2008	32400	0.5993	11340976
0.8513	260.8032	32600	0.6082	11411056
0.5301	262.4016	32800	0.5973	11481152
0.4274	264.0	33000	0.6082	11550752
0.7707	265.6024	33200	0.6110	11620752
0.5863	267.2008	33400	0.6022	11690464
0.6638	268.8032	33600	0.6062	11761360
0.8022	270.4016	33800	0.6082	11831152
0.4962	272.0	34000	0.6052	11900768
0.7421	273.6024	34200	0.6155	11971616
0.8621	275.2008	34400	0.6042	12041104
0.4739	276.8032	34600	0.6042	12111712
0.661	278.4016	34800	0.6115	12181328
0.5588	280.0	35000	0.6040	12251088
0.8743	281.6024	35200	0.6042	12321616
0.5744	283.2008	35400	0.6042	12391184
0.6344	284.8032	35600	0.6042	12461088
0.7548	286.4016	35800	0.6042	12531520
1.0844	288.0	36000	0.6042	12600944
0.3644	289.6024	36200	0.6042	12670544
0.7256	291.2008	36400	0.6042	12741216
0.8211	292.8032	36600	0.6042	12811584
0.6064	294.4016	36800	0.6042	12881104
0.5569	296.0	37000	0.6042	12951648
0.5618	297.6024	37200	0.6042	13021600
0.6211	299.2008	37400	0.6042	13091888
0.5256	300.8032	37600	0.6042	13162128
1.1123	302.4016	37800	0.6042	13231552
0.7682	304.0	38000	0.6042	13302080
0.6204	305.6024	38200	0.6042	13371808
0.8488	307.2008	38400	0.6042	13441936
0.947	308.8032	38600	0.6042	13512304
0.8	310.4016	38800	0.6042	13582192
0.802	312.0	39000	0.6042	13652384
0.457	313.6024	39200	0.6042	13722224
0.6368	315.2008	39400	0.6042	13791728
0.5913	316.8032	39600	0.6042	13862560
0.6218	318.4016	39800	0.6042	13933264
0.6923	320.0	40000	0.6042	14002704

Framework versions

• PEFT 0.15.2.dev0
• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1