You can deploy this model and start earning money today!

RichardErkhov/
jkazdan_-_llama8b-gsm-real-and-synthetic-sftsd2-gguf

Name: jkazdan_-_llama8b-gsm-real-and-synthetic-sftsd2-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

llama8b-gsm-real-and-synthetic-sftsd2 - GGUF

Model creator: https://huggingface.co/jkazdan/
Original model: https://huggingface.co/jkazdan/llama8b-gsm-real-and-synthetic-sftsd2/

Name	Quant method	Size
llama8b-gsm-real-and-synthetic-sftsd2.Q2_K.gguf	Q2_K	2.96GB
llama8b-gsm-real-and-synthetic-sftsd2.IQ3_XS.gguf	IQ3_XS	3.28GB
llama8b-gsm-real-and-synthetic-sftsd2.IQ3_S.gguf	IQ3_S	3.43GB
llama8b-gsm-real-and-synthetic-sftsd2.Q3_K_S.gguf	Q3_K_S	3.41GB
llama8b-gsm-real-and-synthetic-sftsd2.IQ3_M.gguf	IQ3_M	3.52GB
llama8b-gsm-real-and-synthetic-sftsd2.Q3_K.gguf	Q3_K	3.74GB
llama8b-gsm-real-and-synthetic-sftsd2.Q3_K_M.gguf	Q3_K_M	3.74GB
llama8b-gsm-real-and-synthetic-sftsd2.Q3_K_L.gguf	Q3_K_L	4.03GB
llama8b-gsm-real-and-synthetic-sftsd2.IQ4_XS.gguf	IQ4_XS	4.18GB
llama8b-gsm-real-and-synthetic-sftsd2.Q4_0.gguf	Q4_0	4.34GB
llama8b-gsm-real-and-synthetic-sftsd2.IQ4_NL.gguf	IQ4_NL	4.38GB
llama8b-gsm-real-and-synthetic-sftsd2.Q4_K_S.gguf	Q4_K_S	4.37GB
llama8b-gsm-real-and-synthetic-sftsd2.Q4_K.gguf	Q4_K	4.58GB
llama8b-gsm-real-and-synthetic-sftsd2.Q4_K_M.gguf	Q4_K_M	4.58GB
llama8b-gsm-real-and-synthetic-sftsd2.Q4_1.gguf	Q4_1	4.78GB
llama8b-gsm-real-and-synthetic-sftsd2.Q5_0.gguf	Q5_0	5.21GB
llama8b-gsm-real-and-synthetic-sftsd2.Q5_K_S.gguf	Q5_K_S	5.21GB
llama8b-gsm-real-and-synthetic-sftsd2.Q5_K.gguf	Q5_K	5.34GB
llama8b-gsm-real-and-synthetic-sftsd2.Q5_K_M.gguf	Q5_K_M	5.34GB
llama8b-gsm-real-and-synthetic-sftsd2.Q5_1.gguf	Q5_1	5.65GB
llama8b-gsm-real-and-synthetic-sftsd2.Q6_K.gguf	Q6_K	6.14GB
llama8b-gsm-real-and-synthetic-sftsd2.Q8_0.gguf	Q8_0	7.95GB

Original model description:

library_name: transformers license: llama3 base_model: meta-llama/Meta-Llama-3-8B-Instruct tags:

trl
sft
generated_from_trainer model-index:
name: llama8b-gsm-real-and-synthetic-sftsd2 results: []

llama8b-gsm-real-and-synthetic-sftsd2

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

Loss: 1.0859
Num Input Tokens Seen: 1871590

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: 8e-06
train_batch_size: 2
eval_batch_size: 2
seed: 2
gradient_accumulation_steps: 16
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: constant_with_warmup
lr_scheduler_warmup_ratio: 0.05
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
No log	0	0	1.8595	0
2.1954	0.0109	5	1.7942	20052
1.7568	0.0218	10	1.5596	39800
1.4222	0.0327	15	1.3754	61596
1.2425	0.0435	20	1.2578	83300
1.245	0.0544	25	1.2119	103370
1.2296	0.0653	30	1.2007	122848
1.2248	0.0762	35	1.1861	141818
1.2332	0.0871	40	1.1735	163504
1.2422	0.0980	45	1.1752	185520
1.0799	0.1089	50	1.1702	204314
1.1989	0.1198	55	1.1578	225444
1.059	0.1306	60	1.1575	246092
1.1498	0.1415	65	1.1507	267342
1.162	0.1524	70	1.1482	287608
1.2161	0.1633	75	1.1499	305738
1.152	0.1742	80	1.1435	325622
1.1986	0.1851	85	1.1413	346692
1.1673	0.1960	90	1.1410	366070
1.1269	0.2069	95	1.1353	387020
1.08	0.2177	100	1.1345	408372
1.156	0.2286	105	1.1357	427694
1.1383	0.2395	110	1.1358	446868
1.1662	0.2504	115	1.1330	467136
1.1412	0.2613	120	1.1285	489648
1.2021	0.2722	125	1.1302	508680
1.1991	0.2831	130	1.1268	528950
1.1486	0.2940	135	1.1246	550872
1.1836	0.3048	140	1.1271	570586
1.2152	0.3157	145	1.1242	591580
1.2335	0.3266	150	1.1225	611042
1.1121	0.3375	155	1.1205	631400
1.2184	0.3484	160	1.1223	650180
1.168	0.3593	165	1.1189	669268
1.0356	0.3702	170	1.1207	689160
1.1695	0.3811	175	1.1166	709642
1.2066	0.3919	180	1.1150	732480
1.0662	0.4028	185	1.1146	754606
1.1363	0.4137	190	1.1141	775696
1.1564	0.4246	195	1.1128	795878
1.1145	0.4355	200	1.1122	813272
1.269	0.4464	205	1.1137	833902
1.1584	0.4573	210	1.1110	852516
1.16	0.4682	215	1.1096	873596
1.2247	0.4790	220	1.1107	894704
1.0643	0.4899	225	1.1068	914992
1.3557	0.5008	230	1.1081	935502
1.1839	0.5117	235	1.1096	956256
1.1503	0.5226	240	1.1039	977604
1.1692	0.5335	245	1.1043	998968
1.1298	0.5444	250	1.1034	1020772
1.1325	0.5553	255	1.1035	1041154
1.1725	0.5661	260	1.1072	1059292
1.0728	0.5770	265	1.1031	1081362
1.1917	0.5879	270	1.1007	1101692
1.0961	0.5988	275	1.1027	1121708
1.1835	0.6097	280	1.0997	1141298
1.13	0.6206	285	1.0996	1162730
1.1354	0.6315	290	1.1004	1182310
1.1653	0.6424	295	1.1001	1201826
1.0729	0.6532	300	1.0999	1223770
1.1693	0.6641	305	1.0971	1243196
1.1165	0.6750	310	1.0962	1265634
1.0549	0.6859	315	1.0965	1287874
1.0439	0.6968	320	1.0971	1309404
1.1307	0.7077	325	1.0959	1329792
1.0235	0.7186	330	1.0940	1349982
1.1361	0.7295	335	1.0941	1371082
1.1172	0.7403	340	1.0956	1391146
1.077	0.7512	345	1.0931	1412716
1.0474	0.7621	350	1.0928	1433118
1.1478	0.7730	355	1.0932	1453388
1.2037	0.7839	360	1.0916	1475354
1.1049	0.7948	365	1.0925	1495380
1.1686	0.8057	370	1.0929	1515826
1.1108	0.8165	375	1.0920	1534992
1.1364	0.8274	380	1.0900	1555420
1.0312	0.8383	385	1.0893	1576582
1.1645	0.8492	390	1.0903	1596630
1.0845	0.8601	395	1.0897	1616488
1.0322	0.8710	400	1.0899	1636066
1.1525	0.8819	405	1.0898	1655024
1.0964	0.8928	410	1.0889	1674606
1.1863	0.9036	415	1.0870	1693680
1.1249	0.9145	420	1.0880	1712584
1.0701	0.9254	425	1.0876	1734126
1.1546	0.9363	430	1.0859	1754370
1.1891	0.9472	435	1.0884	1773860
1.1046	0.9581	440	1.0861	1795494
1.1069	0.9690	445	1.0840	1814092
1.0491	0.9799	450	1.0863	1834998
1.0807	0.9907	455	1.0857	1855086

Framework versions

Transformers 4.46.0
Pytorch 2.4.1.post300
Datasets 2.20.0
Tokenizers 0.20.1

RichardErkhov/jkazdan_-_llama8b-gsm-real-and-synthetic-sftsd2-gguf

Original model description:

llama8b-gsm-real-and-synthetic-sftsd2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

RichardErkhov/
jkazdan_-_llama8b-gsm-real-and-synthetic-sftsd2-gguf