You can deploy this model and start earning money today!

RichardErkhov/
tsavage68_-_IE_M2_1000steps_1e6rate_SFT-gguf

Name: tsavage68_-_IE_M2_1000steps_1e6rate_SFT-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

IE_M2_1000steps_1e6rate_SFT - GGUF

Model creator: https://huggingface.co/tsavage68/
Original model: https://huggingface.co/tsavage68/IE_M2_1000steps_1e6rate_SFT/

Name	Quant method	Size
IE_M2_1000steps_1e6rate_SFT.Q2_K.gguf	Q2_K	2.53GB
IE_M2_1000steps_1e6rate_SFT.IQ3_XS.gguf	IQ3_XS	2.81GB
IE_M2_1000steps_1e6rate_SFT.IQ3_S.gguf	IQ3_S	2.96GB
IE_M2_1000steps_1e6rate_SFT.Q3_K_S.gguf	Q3_K_S	2.95GB
IE_M2_1000steps_1e6rate_SFT.IQ3_M.gguf	IQ3_M	3.06GB
IE_M2_1000steps_1e6rate_SFT.Q3_K.gguf	Q3_K	3.28GB
IE_M2_1000steps_1e6rate_SFT.Q3_K_M.gguf	Q3_K_M	3.28GB
IE_M2_1000steps_1e6rate_SFT.Q3_K_L.gguf	Q3_K_L	3.56GB
IE_M2_1000steps_1e6rate_SFT.IQ4_XS.gguf	IQ4_XS	3.67GB
IE_M2_1000steps_1e6rate_SFT.Q4_0.gguf	Q4_0	3.83GB
IE_M2_1000steps_1e6rate_SFT.IQ4_NL.gguf	IQ4_NL	3.87GB
IE_M2_1000steps_1e6rate_SFT.Q4_K_S.gguf	Q4_K_S	3.86GB
IE_M2_1000steps_1e6rate_SFT.Q4_K.gguf	Q4_K	4.07GB
IE_M2_1000steps_1e6rate_SFT.Q4_K_M.gguf	Q4_K_M	4.07GB
IE_M2_1000steps_1e6rate_SFT.Q4_1.gguf	Q4_1	4.24GB
IE_M2_1000steps_1e6rate_SFT.Q5_0.gguf	Q5_0	4.65GB
IE_M2_1000steps_1e6rate_SFT.Q5_K_S.gguf	Q5_K_S	4.65GB
IE_M2_1000steps_1e6rate_SFT.Q5_K.gguf	Q5_K	4.78GB
IE_M2_1000steps_1e6rate_SFT.Q5_K_M.gguf	Q5_K_M	4.78GB
IE_M2_1000steps_1e6rate_SFT.Q5_1.gguf	Q5_1	5.07GB
IE_M2_1000steps_1e6rate_SFT.Q6_K.gguf	Q6_K	5.53GB
IE_M2_1000steps_1e6rate_SFT.Q8_0.gguf	Q8_0	7.17GB

Original model description:

library_name: transformers license: apache-2.0 base_model: mistralai/Mistral-7B-Instruct-v0.2 tags:

trl
sft
generated_from_trainer model-index:
name: IE_M2_1000steps_1e6rate_SFT results: []

IE_M2_1000steps_1e6rate_SFT

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 2.1241

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: 1e-06
train_batch_size: 2
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss
1.3204	0.4	50	1.2951
1.2959	0.8	100	1.2707
1.1469	1.2	150	1.2966
1.0839	1.6	200	1.2918
1.1089	2.0	250	1.2759
0.6433	2.4	300	1.3874
0.6433	2.8	350	1.3916
0.3106	3.2	400	1.6073
0.3397	3.6	450	1.5376
0.2909	4.0	500	1.6111
0.1419	4.4	550	1.7819
0.1438	4.8	600	1.7870
0.0583	5.2	650	1.9767
0.0645	5.6	700	1.9910
0.0634	6.0	750	1.9911
0.0383	6.4	800	2.1028
0.0405	6.8	850	2.1131
0.0292	7.2	900	2.1218
0.0284	7.6	950	2.1242
0.0323	8.0	1000	2.1241

Framework versions

Transformers 4.44.2
Pytorch 2.0.0+cu117
Datasets 3.0.0
Tokenizers 0.19.1

RichardErkhov/tsavage68_-_IE_M2_1000steps_1e6rate_SFT-gguf

Original model description:

IE_M2_1000steps_1e6rate_SFT

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

RichardErkhov/
tsavage68_-_IE_M2_1000steps_1e6rate_SFT-gguf