train_wsc_1745950301

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: 0.3479
• 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: 0.3
• 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
0.3481	1.6024	200	0.3937	70144
0.3618	3.2008	400	0.3625	140304
0.3966	4.8032	600	0.3609	210240
0.3759	6.4016	800	0.4168	279952
0.5142	8.0	1000	0.3932	350224
0.3172	9.6024	1200	0.4967	420256
0.3539	11.2008	1400	0.6324	490496
0.3909	12.8032	1600	0.3521	560224
0.3751	14.4016	1800	0.3479	630560
0.357	16.0	2000	0.3643	699648
0.3893	17.6024	2200	0.3549	769232
0.3175	19.2008	2400	0.4833	839344
0.3652	20.8032	2600	0.3520	909744
0.365	22.4016	2800	0.3521	979312
0.3945	24.0	3000	0.3519	1049184
0.3726	25.6024	3200	0.3594	1119552
0.3951	27.2008	3400	0.3498	1189008
0.3497	28.8032	3600	0.3815	1259168
0.3087	30.4016	3800	0.3790	1329056
0.3478	32.0	4000	0.3681	1399280
0.3321	33.6024	4200	0.4623	1469920
0.3297	35.2008	4400	0.3859	1539184
0.3218	36.8032	4600	0.4085	1609648
0.2996	38.4016	4800	0.4424	1679792
0.4013	40.0	5000	0.3618	1749008
0.368	41.6024	5200	0.3772	1818832
0.3804	43.2008	5400	0.3532	1889136
0.3447	44.8032	5600	0.3504	1959008
0.4024	46.4016	5800	0.3740	2028320
0.3575	48.0	6000	0.3546	2098928
0.3726	49.6024	6200	0.3559	2168688
0.3459	51.2008	6400	0.3536	2238752
0.3578	52.8032	6600	0.3571	2308816
0.3395	54.4016	6800	0.3686	2379328
0.3692	56.0	7000	0.3688	2448704
0.5154	57.6024	7200	0.3540	2519008
0.3707	59.2008	7400	0.3510	2588608
0.3494	60.8032	7600	0.3638	2659072
0.3521	62.4016	7800	0.3524	2728480
0.4449	64.0	8000	0.3593	2798720
0.3794	65.6024	8200	0.3858	2868672
0.3643	67.2008	8400	0.3597	2939312
0.3434	68.8032	8600	0.3513	3009568
0.3494	70.4016	8800	0.3696	3079584
0.3478	72.0	9000	0.3524	3149680
0.3234	73.6024	9200	0.4030	3219680
0.3491	75.2008	9400	0.3532	3289472
0.3474	76.8032	9600	0.3538	3359520
0.3429	78.4016	9800	0.3582	3429568
0.3524	80.0	10000	0.3500	3499648
0.3272	81.6024	10200	0.3656	3569504
0.3907	83.2008	10400	0.3989	3639920
0.2551	84.8032	10600	0.4358	3709520
0.372	86.4016	10800	0.3547	3779456
0.3645	88.0	11000	0.3545	3849744
0.384	89.6024	11200	0.3532	3919984
0.3421	91.2008	11400	0.3520	3989872
0.3697	92.8032	11600	0.3584	4059568
0.3618	94.4016	11800	0.3497	4129664
0.3462	96.0	12000	0.3715	4199936
0.3189	97.6024	12200	0.3875	4269952
0.3483	99.2008	12400	0.3619	4339040
0.3477	100.8032	12600	0.3564	4409680
0.3459	102.4016	12800	0.3587	4479120
0.3518	104.0	13000	0.4024	4548896
0.3558	105.6024	13200	0.3599	4619216
0.3899	107.2008	13400	0.3608	4689424
0.375	108.8032	13600	0.3554	4759232
0.3441	110.4016	13800	0.3636	4829120
0.3495	112.0	14000	0.3556	4899024
0.3535	113.6024	14200	0.3591	4968944
0.3393	115.2008	14400	0.3589	5039152
0.3857	116.8032	14600	0.3566	5109312
0.345	118.4016	14800	0.3546	5179296
0.351	120.0	15000	0.3538	5249504
0.3259	121.6024	15200	0.3612	5319424
0.3209	123.2008	15400	0.3808	5389488
0.3565	124.8032	15600	0.3535	5459776
0.3271	126.4016	15800	0.3515	5529760
0.3092	128.0	16000	0.3808	5599968
0.3434	129.6024	16200	0.3500	5671056
0.3532	131.2008	16400	0.3604	5740000
0.3681	132.8032	16600	0.3572	5810288
0.353	134.4016	16800	0.3594	5880176
0.3471	136.0	17000	0.3579	5950048
0.3562	137.6024	17200	0.3644	6020016
0.3892	139.2008	17400	0.3583	6090672
0.3545	140.8032	17600	0.3681	6160288
0.4053	142.4016	17800	0.3721	6230656
0.3224	144.0	18000	0.3567	6299968
0.3377	145.6024	18200	0.3646	6370512
0.3491	147.2008	18400	0.3558	6440784
0.3411	148.8032	18600	0.3606	6510560
0.3344	150.4016	18800	0.3552	6579872
0.3227	152.0	19000	0.3651	6650112
0.3469	153.6024	19200	0.3702	6720368
0.3872	155.2008	19400	0.3737	6790512
0.3488	156.8032	19600	0.3525	6860880
0.3635	158.4016	19800	0.3770	6930576
0.34	160.0	20000	0.3582	7000640
0.3565	161.6024	20200	0.3523	7070272
0.3411	163.2008	20400	0.3561	7140336
0.3373	164.8032	20600	0.3497	7210816
0.3482	166.4016	20800	0.3670	7281392
0.339	168.0	21000	0.3549	7350960
0.3145	169.6024	21200	0.3669	7421312
0.3461	171.2008	21400	0.3559	7491200
0.3472	172.8032	21600	0.3576	7560976
0.3532	174.4016	21800	0.3503	7631024
0.3441	176.0	22000	0.3551	7700784
0.3545	177.6024	22200	0.3680	7770752
0.4	179.2008	22400	0.3657	7840832
0.3275	180.8032	22600	0.3675	7911072
0.3382	182.4016	22800	0.3553	7981312
0.3682	184.0	23000	0.3611	8050976
0.2797	185.6024	23200	0.3805	8121312
0.3475	187.2008	23400	0.3546	8191520
0.3506	188.8032	23600	0.3532	8261456
0.3341	190.4016	23800	0.3702	8331664
0.328	192.0	24000	0.3560	8401328
0.3563	193.6024	24200	0.3561	8471232
0.3585	195.2008	24400	0.3580	8540976
0.3998	196.8032	24600	0.3776	8611296
0.3351	198.4016	24800	0.3581	8681264
0.3714	200.0	25000	0.3618	8751280
0.35	201.6024	25200	0.3553	8822192
0.3299	203.2008	25400	0.3635	8891648
0.3368	204.8032	25600	0.3604	8961760
0.3453	206.4016	25800	0.3571	9031568
0.3574	208.0	26000	0.3588	9101088
0.3359	209.6024	26200	0.3531	9171168
0.3649	211.2008	26400	0.3597	9240752
0.3464	212.8032	26600	0.3524	9310960
0.3582	214.4016	26800	0.3685	9380560
0.3518	216.0	27000	0.3577	9450912
0.3405	217.6024	27200	0.3542	9520832
0.3337	219.2008	27400	0.3536	9590800
0.3373	220.8032	27600	0.3539	9661456
0.3101	222.4016	27800	0.3652	9731376
0.3749	224.0	28000	0.3654	9801040
0.3415	225.6024	28200	0.3558	9870784
0.3449	227.2008	28400	0.3590	9941408
0.328	228.8032	28600	0.3614	10011264
0.3322	230.4016	28800	0.3608	10080704
0.3209	232.0	29000	0.3612	10150880
0.3315	233.6024	29200	0.3677	10221616
0.3314	235.2008	29400	0.3679	10291664
0.3386	236.8032	29600	0.3543	10361728
0.347	238.4016	29800	0.3540	10431088
0.3694	240.0	30000	0.3702	10501088
0.3238	241.6024	30200	0.3639	10571488
0.3311	243.2008	30400	0.3622	10640848
0.3445	244.8032	30600	0.3631	10711136
0.3558	246.4016	30800	0.3615	10781136
0.3495	248.0	31000	0.3610	10851312
0.361	249.6024	31200	0.3544	10921664
0.3543	251.2008	31400	0.3628	10991936
0.351	252.8032	31600	0.3619	11061680
0.3288	254.4016	31800	0.3700	11131872
0.3503	256.0	32000	0.3581	11201520
0.3545	257.6024	32200	0.3688	11271952
0.3452	259.2008	32400	0.3665	11340976
0.3451	260.8032	32600	0.3572	11411056
0.3492	262.4016	32800	0.3594	11481152
0.37	264.0	33000	0.3602	11550752
0.3444	265.6024	33200	0.3605	11620752
0.3474	267.2008	33400	0.3590	11690464
0.3421	268.8032	33600	0.3647	11761360
0.3466	270.4016	33800	0.3618	11831152
0.3418	272.0	34000	0.3609	11900768
0.3394	273.6024	34200	0.3612	11971616
0.3319	275.2008	34400	0.3632	12041104
0.3679	276.8032	34600	0.3596	12111712
0.3522	278.4016	34800	0.3598	12181328
0.3434	280.0	35000	0.3597	12251088
0.3281	281.6024	35200	0.3560	12321616
0.3377	283.2008	35400	0.3551	12391184
0.3346	284.8032	35600	0.3605	12461088
0.3374	286.4016	35800	0.3595	12531520
0.3407	288.0	36000	0.3593	12600944
0.362	289.6024	36200	0.3630	12670544
0.3365	291.2008	36400	0.3603	12741216
0.3319	292.8032	36600	0.3668	12811584
0.3266	294.4016	36800	0.3617	12881104
0.3582	296.0	37000	0.3609	12951648
0.3432	297.6024	37200	0.3629	13021600
0.342	299.2008	37400	0.3624	13091888
0.3658	300.8032	37600	0.3633	13162128
0.3142	302.4016	37800	0.3627	13231552
0.331	304.0	38000	0.3613	13302080
0.3507	305.6024	38200	0.3595	13371808
0.3403	307.2008	38400	0.3596	13441936
0.3275	308.8032	38600	0.3583	13512304
0.3553	310.4016	38800	0.3591	13582192
0.3348	312.0	39000	0.3615	13652384
0.3715	313.6024	39200	0.3620	13722224
0.3552	315.2008	39400	0.3578	13791728
0.3445	316.8032	39600	0.3609	13862560
0.3485	318.4016	39800	0.3606	13933264
0.3448	320.0	40000	0.3591	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