Hugging Face's logo

Jsoo
/
results_1011

PEFT
TensorBoard
Safetensors
trl
sft
Generated from Trainer
Model card Files Metrics Community
Jsoo commited on
Commit
aa89300
•
1 Parent(s): 2210c5e

Jsoo/llama3.2-3b-hard

Browse files
Files changed (5) hide show
  1. README.md +29 -129
  2. adapter_model.safetensors +1 -1
  3. runs/Oct23_12-48-29_DESKTOP-UVSTCBR/events.out.tfevents.1729655317.DESKTOP-UVSTCBR.210884.0 +3 -0
  4. tokenizer.json +2 -2
  5. training_args.bin +1 -1
README.md CHANGED
@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [meta-llama/Llama-3.2-3B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct) on the None dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 1.7021
22
 
23
  ## Model description
24
 
@@ -46,138 +46,38 @@ The following hyperparameters were used during training:
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: cosine
48
  - lr_scheduler_warmup_ratio: 0.01
49
- - num_epochs: 10
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
- | Training Loss | Epoch | Step | Validation Loss |
55
- |:-------------:|:------:|:-----:|:---------------:|
56
- | 2.8038 | 0.08 | 100 | 2.5553 |
57
- | 2.507 | 0.16 | 200 | 2.4913 |
58
- | 2.4522 | 0.24 | 300 | 2.4433 |
59
- | 2.417 | 0.32 | 400 | 2.4122 |
60
- | 2.3949 | 0.4 | 500 | 2.3822 |
61
- | 2.3586 | 0.48 | 600 | 2.3510 |
62
- | 2.3359 | 0.56 | 700 | 2.3276 |
63
- | 2.3081 | 0.64 | 800 | 2.3043 |
64
- | 2.307 | 0.72 | 900 | 2.2872 |
65
- | 2.2741 | 0.8 | 1000 | 2.2695 |
66
- | 2.2594 | 0.88 | 1100 | 2.2515 |
67
- | 2.2591 | 0.96 | 1200 | 2.2363 |
68
- | 2.2163 | 1.04 | 1300 | 2.2191 |
69
- | 2.1996 | 1.12 | 1400 | 2.2057 |
70
- | 2.1729 | 1.2 | 1500 | 2.1928 |
71
- | 2.1718 | 1.28 | 1600 | 2.1799 |
72
- | 2.1586 | 1.3600 | 1700 | 2.1692 |
73
- | 2.1386 | 1.44 | 1800 | 2.1530 |
74
- | 2.1338 | 1.52 | 1900 | 2.1418 |
75
- | 2.1189 | 1.6 | 2000 | 2.1307 |
76
- | 2.1055 | 1.6800 | 2100 | 2.1187 |
77
- | 2.1074 | 1.76 | 2200 | 2.1075 |
78
- | 2.0919 | 1.8400 | 2300 | 2.0959 |
79
- | 2.0812 | 1.92 | 2400 | 2.0845 |
80
- | 2.0621 | 2.0 | 2500 | 2.0743 |
81
- | 2.0172 | 2.08 | 2600 | 2.0666 |
82
- | 2.0159 | 2.16 | 2700 | 2.0602 |
83
- | 2.0075 | 2.24 | 2800 | 2.0476 |
84
- | 2.0042 | 2.32 | 2900 | 2.0394 |
85
- | 2.0062 | 2.4 | 3000 | 2.0262 |
86
- | 1.989 | 2.48 | 3100 | 2.0157 |
87
- | 1.9808 | 2.56 | 3200 | 2.0086 |
88
- | 1.9792 | 2.64 | 3300 | 1.9985 |
89
- | 1.9751 | 2.7200 | 3400 | 1.9904 |
90
- | 1.963 | 2.8 | 3500 | 1.9810 |
91
- | 1.9498 | 2.88 | 3600 | 1.9718 |
92
- | 1.9495 | 2.96 | 3700 | 1.9662 |
93
- | 1.9053 | 3.04 | 3800 | 1.9578 |
94
- | 1.8905 | 3.12 | 3900 | 1.9486 |
95
- | 1.8873 | 3.2 | 4000 | 1.9412 |
96
- | 1.8963 | 3.2800 | 4100 | 1.9347 |
97
- | 1.8847 | 3.36 | 4200 | 1.9274 |
98
- | 1.8819 | 3.44 | 4300 | 1.9187 |
99
- | 1.8789 | 3.52 | 4400 | 1.9151 |
100
- | 1.8635 | 3.6 | 4500 | 1.9057 |
101
- | 1.8557 | 3.68 | 4600 | 1.9010 |
102
- | 1.8518 | 3.76 | 4700 | 1.8927 |
103
- | 1.8444 | 3.84 | 4800 | 1.8863 |
104
- | 1.8318 | 3.92 | 4900 | 1.8801 |
105
- | 1.8387 | 4.0 | 5000 | 1.8737 |
106
- | 1.7994 | 4.08 | 5100 | 1.8701 |
107
- | 1.7866 | 4.16 | 5200 | 1.8634 |
108
- | 1.8005 | 4.24 | 5300 | 1.8623 |
109
- | 1.7951 | 4.32 | 5400 | 1.8558 |
110
- | 1.7818 | 4.4 | 5500 | 1.8477 |
111
- | 1.7874 | 4.48 | 5600 | 1.8426 |
112
- | 1.7771 | 4.5600 | 5700 | 1.8386 |
113
- | 1.7574 | 4.64 | 5800 | 1.8353 |
114
- | 1.7758 | 4.72 | 5900 | 1.8273 |
115
- | 1.7864 | 4.8 | 6000 | 1.8244 |
116
- | 1.7741 | 4.88 | 6100 | 1.8262 |
117
- | 1.7638 | 4.96 | 6200 | 1.8151 |
118
- | 1.7485 | 5.04 | 6300 | 1.8085 |
119
- | 1.7239 | 5.12 | 6400 | 1.8017 |
120
- | 1.7231 | 5.2 | 6500 | 1.7985 |
121
- | 1.7212 | 5.28 | 6600 | 1.7950 |
122
- | 1.7183 | 5.36 | 6700 | 1.7907 |
123
- | 1.7234 | 5.44 | 6800 | 1.7856 |
124
- | 1.7082 | 5.52 | 6900 | 1.7830 |
125
- | 1.7128 | 5.6 | 7000 | 1.7792 |
126
- | 1.7114 | 5.68 | 7100 | 1.7743 |
127
- | 1.7193 | 5.76 | 7200 | 1.7714 |
128
- | 1.7093 | 5.84 | 7300 | 1.7672 |
129
- | 1.6974 | 5.92 | 7400 | 1.7643 |
130
- | 1.7176 | 6.0 | 7500 | 1.7599 |
131
- | 1.6657 | 6.08 | 7600 | 1.7575 |
132
- | 1.679 | 6.16 | 7700 | 1.7560 |
133
- | 1.6663 | 6.24 | 7800 | 1.7526 |
134
- | 1.6634 | 6.32 | 7900 | 1.7499 |
135
- | 1.6736 | 6.4 | 8000 | 1.7466 |
136
- | 1.661 | 6.48 | 8100 | 1.7448 |
137
- | 1.6535 | 6.5600 | 8200 | 1.7438 |
138
- | 1.6734 | 6.64 | 8300 | 1.7395 |
139
- | 1.6611 | 6.72 | 8400 | 1.7370 |
140
- | 1.6841 | 6.8 | 8500 | 1.7337 |
141
- | 1.6735 | 6.88 | 8600 | 1.7331 |
142
- | 1.6679 | 6.96 | 8700 | 1.7316 |
143
- | 1.6459 | 7.04 | 8800 | 1.7305 |
144
- | 1.6438 | 7.12 | 8900 | 1.7296 |
145
- | 1.6436 | 7.2 | 9000 | 1.7283 |
146
- | 1.6293 | 7.28 | 9100 | 1.7278 |
147
- | 1.6424 | 7.36 | 9200 | 1.7252 |
148
- | 1.64 | 7.44 | 9300 | 1.7244 |
149
- | 1.6114 | 7.52 | 9400 | 1.7227 |
150
- | 1.6331 | 7.6 | 9500 | 1.7214 |
151
- | 1.628 | 7.68 | 9600 | 1.7173 |
152
- | 1.6464 | 7.76 | 9700 | 1.7159 |
153
- | 1.6355 | 7.84 | 9800 | 1.7138 |
154
- | 1.6489 | 7.92 | 9900 | 1.7127 |
155
- | 1.6436 | 8.0 | 10000 | 1.7113 |
156
- | 1.6108 | 8.08 | 10100 | 1.7108 |
157
- | 1.6252 | 8.16 | 10200 | 1.7097 |
158
- | 1.6228 | 8.24 | 10300 | 1.7087 |
159
- | 1.617 | 8.32 | 10400 | 1.7084 |
160
- | 1.6255 | 8.4 | 10500 | 1.7079 |
161
- | 1.6212 | 8.48 | 10600 | 1.7070 |
162
- | 1.6146 | 8.56 | 10700 | 1.7068 |
163
- | 1.625 | 8.64 | 10800 | 1.7060 |
164
- | 1.6282 | 8.72 | 10900 | 1.7056 |
165
- | 1.614 | 8.8 | 11000 | 1.7054 |
166
- | 1.612 | 8.88 | 11100 | 1.7051 |
167
- | 1.6145 | 8.96 | 11200 | 1.7040 |
168
- | 1.6125 | 9.04 | 11300 | 1.7037 |
169
- | 1.6282 | 9.12 | 11400 | 1.7030 |
170
- | 1.6085 | 9.2 | 11500 | 1.7030 |
171
- | 1.6008 | 9.28 | 11600 | 1.7027 |
172
- | 1.6109 | 9.36 | 11700 | 1.7024 |
173
- | 1.6318 | 9.44 | 11800 | 1.7023 |
174
- | 1.5976 | 9.52 | 11900 | 1.7022 |
175
- | 1.5975 | 9.6 | 12000 | 1.7022 |
176
- | 1.6108 | 9.68 | 12100 | 1.7021 |
177
- | 1.6158 | 9.76 | 12200 | 1.7021 |
178
- | 1.6232 | 9.84 | 12300 | 1.7021 |
179
- | 1.6109 | 9.92 | 12400 | 1.7021 |
180
- | 1.6005 | 10.0 | 12500 | 1.7021 |
181
 
182
 
183
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [meta-llama/Llama-3.2-3B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct) on the None dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 1.9956
22
 
23
  ## Model description
24
 
 
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: cosine
48
  - lr_scheduler_warmup_ratio: 0.01
49
+ - num_epochs: 20
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
+ | Training Loss | Epoch | Step | Validation Loss |
55
+ |:-------------:|:------:|:----:|:---------------:|
56
+ | 2.6788 | 0.3901 | 100 | 2.2881 |
57
+ | 2.4361 | 0.7801 | 200 | 2.2154 |
58
+ | 2.3903 | 1.1702 | 300 | 2.1747 |
59
+ | 2.3166 | 1.5602 | 400 | 2.1358 |
60
+ | 2.2868 | 1.9503 | 500 | 2.1058 |
61
+ | 2.2048 | 2.3403 | 600 | 2.0800 |
62
+ | 2.1999 | 2.7304 | 700 | 2.0613 |
63
+ | 2.1711 | 3.1204 | 800 | 2.0471 |
64
+ | 2.1038 | 3.5105 | 900 | 2.0329 |
65
+ | 2.1115 | 3.9005 | 1000 | 2.0185 |
66
+ | 2.0859 | 4.2906 | 1100 | 2.0129 |
67
+ | 2.0455 | 4.6806 | 1200 | 2.0084 |
68
+ | 2.0338 | 5.0707 | 1300 | 2.0022 |
69
+ | 1.9991 | 5.4608 | 1400 | 2.0011 |
70
+ | 1.9948 | 5.8508 | 1500 | 1.9966 |
71
+ | 1.948 | 6.2409 | 1600 | 1.9977 |
72
+ | 1.9773 | 6.6309 | 1700 | 1.9909 |
73
+ | 1.9228 | 7.0210 | 1800 | 1.9915 |
74
+ | 1.8997 | 7.4110 | 1900 | 1.9947 |
75
+ | 1.9212 | 7.8011 | 2000 | 1.9868 |
76
+ | 1.8786 | 8.1911 | 2100 | 2.0092 |
77
+ | 1.8762 | 8.5812 | 2200 | 2.0070 |
78
+ | 1.8724 | 8.9712 | 2300 | 2.0023 |
79
+ | 1.8604 | 9.3613 | 2400 | 1.9978 |
80
+ | 1.8436 | 9.7513 | 2500 | 1.9956 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
81
 
82
 
83
  ### Framework versions
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6218ff3ca163c1ffc14aa4de9f6aebbb7b331ae0a484960314076edb87c7199c
3
  size 36715216
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:516752b5acbc965b718d6b574b27a362a8ead0175f19a71abb25be26c449d19a
3
  size 36715216
runs/Oct23_12-48-29_DESKTOP-UVSTCBR/events.out.tfevents.1729655317.DESKTOP-UVSTCBR.210884.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c25b30b39fbe1cc2afddeabd842aa76850ab098ed9317892b72370589074803
3
+ size 17903
tokenizer.json CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:76cfe2f054560aae896b2b75e273dc97a39e304d4ad19c44a9727a1d6b33c4cc
3
- size 17210021
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c85066e7642934ed09b44155e6566b0b5dab2637fb9433439ba5c9c7f8b50d3
3
+ size 17210018
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b80f78d46420cd5b82723bd77d68e989c62705f2fccb568de3478cb149ae85d6
3
  size 5432
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94c489243f58e0385e630c2a899dc27c4bd2d04b61d534b43b69d108a9380236
3
  size 5432