End of training
Browse files- README.md +5 -5
- all_results.json +19 -0
- egy_training_log.txt +2 -0
- eval_results.json +13 -0
- train_results.json +9 -0
- train_vs_val_loss.png +0 -0
- trainer_state.json +184 -0
README.md
CHANGED
@@ -18,11 +18,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
18 |
|
19 |
This model is a fine-tuned version of [riotu-lab/ArabianGPT-03B](https://huggingface.co/riotu-lab/ArabianGPT-03B) on an unknown dataset.
|
20 |
It achieves the following results on the evaluation set:
|
21 |
-
- Loss: 0.
|
22 |
-
- Bleu: 0.
|
23 |
-
- Rouge1: 0.
|
24 |
-
- Rouge2: 0.
|
25 |
-
- Rougel: 0.
|
26 |
|
27 |
## Model description
|
28 |
|
|
|
18 |
|
19 |
This model is a fine-tuned version of [riotu-lab/ArabianGPT-03B](https://huggingface.co/riotu-lab/ArabianGPT-03B) on an unknown dataset.
|
20 |
It achieves the following results on the evaluation set:
|
21 |
+
- Loss: 0.3495
|
22 |
+
- Bleu: 0.4355
|
23 |
+
- Rouge1: 0.6808
|
24 |
+
- Rouge2: 0.4218
|
25 |
+
- Rougel: 0.6799
|
26 |
|
27 |
## Model description
|
28 |
|
all_results.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 7.0,
|
3 |
+
"eval_bleu": 0.43550162080921356,
|
4 |
+
"eval_loss": 0.3495321273803711,
|
5 |
+
"eval_rouge1": 0.6807566585005222,
|
6 |
+
"eval_rouge2": 0.42184236623206983,
|
7 |
+
"eval_rougeL": 0.6798697145255593,
|
8 |
+
"eval_runtime": 135.4262,
|
9 |
+
"eval_samples": 1672,
|
10 |
+
"eval_samples_per_second": 12.346,
|
11 |
+
"eval_steps_per_second": 1.543,
|
12 |
+
"perplexity": 1.4184037610379865,
|
13 |
+
"total_flos": 1.086463725010944e+16,
|
14 |
+
"train_loss": 0.29683067467794466,
|
15 |
+
"train_runtime": 4067.7631,
|
16 |
+
"train_samples": 6685,
|
17 |
+
"train_samples_per_second": 32.868,
|
18 |
+
"train_steps_per_second": 4.11
|
19 |
+
}
|
egy_training_log.txt
CHANGED
@@ -154,3 +154,5 @@ INFO:root:Epoch 6.0: Train Loss = 0.11, Eval Loss = 0.3791240453720093
|
|
154 |
INFO:absl:Using default tokenizer.
|
155 |
INFO:root:Epoch 7.0: Train Loss = 0.0981, Eval Loss = 0.39224255084991455
|
156 |
INFO:absl:Using default tokenizer.
|
|
|
|
|
|
154 |
INFO:absl:Using default tokenizer.
|
155 |
INFO:root:Epoch 7.0: Train Loss = 0.0981, Eval Loss = 0.39224255084991455
|
156 |
INFO:absl:Using default tokenizer.
|
157 |
+
INFO:__main__:*** Evaluate ***
|
158 |
+
INFO:absl:Using default tokenizer.
|
eval_results.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 7.0,
|
3 |
+
"eval_bleu": 0.43550162080921356,
|
4 |
+
"eval_loss": 0.3495321273803711,
|
5 |
+
"eval_rouge1": 0.6807566585005222,
|
6 |
+
"eval_rouge2": 0.42184236623206983,
|
7 |
+
"eval_rougeL": 0.6798697145255593,
|
8 |
+
"eval_runtime": 135.4262,
|
9 |
+
"eval_samples": 1672,
|
10 |
+
"eval_samples_per_second": 12.346,
|
11 |
+
"eval_steps_per_second": 1.543,
|
12 |
+
"perplexity": 1.4184037610379865
|
13 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 7.0,
|
3 |
+
"total_flos": 1.086463725010944e+16,
|
4 |
+
"train_loss": 0.29683067467794466,
|
5 |
+
"train_runtime": 4067.7631,
|
6 |
+
"train_samples": 6685,
|
7 |
+
"train_samples_per_second": 32.868,
|
8 |
+
"train_steps_per_second": 4.11
|
9 |
+
}
|
train_vs_val_loss.png
ADDED
![]() |
trainer_state.json
ADDED
@@ -0,0 +1,184 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.3495321273803711,
|
3 |
+
"best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_gulf_03/checkpoint-1672",
|
4 |
+
"epoch": 7.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 5852,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 1.0,
|
13 |
+
"grad_norm": 1.8581334352493286,
|
14 |
+
"learning_rate": 4.896424167694204e-05,
|
15 |
+
"loss": 1.1759,
|
16 |
+
"step": 836
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 1.0,
|
20 |
+
"eval_bleu": 0.41556477975575595,
|
21 |
+
"eval_loss": 0.37735649943351746,
|
22 |
+
"eval_rouge1": 0.6364308285085805,
|
23 |
+
"eval_rouge2": 0.36296002645709646,
|
24 |
+
"eval_rougeL": 0.6351682159497727,
|
25 |
+
"eval_runtime": 16.4559,
|
26 |
+
"eval_samples_per_second": 101.605,
|
27 |
+
"eval_steps_per_second": 12.701,
|
28 |
+
"step": 836
|
29 |
+
},
|
30 |
+
{
|
31 |
+
"epoch": 2.0,
|
32 |
+
"grad_norm": 2.0768027305603027,
|
33 |
+
"learning_rate": 4.638717632552405e-05,
|
34 |
+
"loss": 0.2824,
|
35 |
+
"step": 1672
|
36 |
+
},
|
37 |
+
{
|
38 |
+
"epoch": 2.0,
|
39 |
+
"eval_bleu": 0.43550162080921356,
|
40 |
+
"eval_loss": 0.3495321273803711,
|
41 |
+
"eval_rouge1": 0.6807566585005222,
|
42 |
+
"eval_rouge2": 0.42184236623206983,
|
43 |
+
"eval_rougeL": 0.6798697145255593,
|
44 |
+
"eval_runtime": 12.3396,
|
45 |
+
"eval_samples_per_second": 135.499,
|
46 |
+
"eval_steps_per_second": 16.937,
|
47 |
+
"step": 1672
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 3.0,
|
51 |
+
"grad_norm": 1.4890309572219849,
|
52 |
+
"learning_rate": 4.3810110974106046e-05,
|
53 |
+
"loss": 0.1846,
|
54 |
+
"step": 2508
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 3.0,
|
58 |
+
"eval_bleu": 0.44798984557044996,
|
59 |
+
"eval_loss": 0.35791030526161194,
|
60 |
+
"eval_rouge1": 0.6988772317351744,
|
61 |
+
"eval_rouge2": 0.4541195944585126,
|
62 |
+
"eval_rougeL": 0.6983599524493045,
|
63 |
+
"eval_runtime": 13.092,
|
64 |
+
"eval_samples_per_second": 127.712,
|
65 |
+
"eval_steps_per_second": 15.964,
|
66 |
+
"step": 2508
|
67 |
+
},
|
68 |
+
{
|
69 |
+
"epoch": 4.0,
|
70 |
+
"grad_norm": 0.9949209690093994,
|
71 |
+
"learning_rate": 4.1233045622688044e-05,
|
72 |
+
"loss": 0.1351,
|
73 |
+
"step": 3344
|
74 |
+
},
|
75 |
+
{
|
76 |
+
"epoch": 4.0,
|
77 |
+
"eval_bleu": 0.4449484201171957,
|
78 |
+
"eval_loss": 0.36742404103279114,
|
79 |
+
"eval_rouge1": 0.6992663461784911,
|
80 |
+
"eval_rouge2": 0.45557653256736347,
|
81 |
+
"eval_rougeL": 0.6984837283971637,
|
82 |
+
"eval_runtime": 12.3128,
|
83 |
+
"eval_samples_per_second": 135.794,
|
84 |
+
"eval_steps_per_second": 16.974,
|
85 |
+
"step": 3344
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 5.0,
|
89 |
+
"grad_norm": 1.245552897453308,
|
90 |
+
"learning_rate": 3.8655980271270036e-05,
|
91 |
+
"loss": 0.11,
|
92 |
+
"step": 4180
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 5.0,
|
96 |
+
"eval_bleu": 0.4488174409372482,
|
97 |
+
"eval_loss": 0.3791240453720093,
|
98 |
+
"eval_rouge1": 0.7066508462378398,
|
99 |
+
"eval_rouge2": 0.46813923665655116,
|
100 |
+
"eval_rougeL": 0.7055578408590141,
|
101 |
+
"eval_runtime": 18.2521,
|
102 |
+
"eval_samples_per_second": 91.606,
|
103 |
+
"eval_steps_per_second": 11.451,
|
104 |
+
"step": 4180
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 6.0,
|
108 |
+
"grad_norm": 1.253645896911621,
|
109 |
+
"learning_rate": 3.6078914919852034e-05,
|
110 |
+
"loss": 0.0981,
|
111 |
+
"step": 5016
|
112 |
+
},
|
113 |
+
{
|
114 |
+
"epoch": 6.0,
|
115 |
+
"eval_bleu": 0.4582553582864403,
|
116 |
+
"eval_loss": 0.39224255084991455,
|
117 |
+
"eval_rouge1": 0.7135223253955782,
|
118 |
+
"eval_rouge2": 0.4789126140106226,
|
119 |
+
"eval_rougeL": 0.7129027733719826,
|
120 |
+
"eval_runtime": 65.9112,
|
121 |
+
"eval_samples_per_second": 25.367,
|
122 |
+
"eval_steps_per_second": 3.171,
|
123 |
+
"step": 5016
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 7.0,
|
127 |
+
"grad_norm": 0.6491497755050659,
|
128 |
+
"learning_rate": 3.350184956843403e-05,
|
129 |
+
"loss": 0.0917,
|
130 |
+
"step": 5852
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 7.0,
|
134 |
+
"eval_bleu": 0.4567065309085701,
|
135 |
+
"eval_loss": 0.3964388072490692,
|
136 |
+
"eval_rouge1": 0.7126519996998933,
|
137 |
+
"eval_rouge2": 0.47787123299802714,
|
138 |
+
"eval_rougeL": 0.7120354739021442,
|
139 |
+
"eval_runtime": 21.7898,
|
140 |
+
"eval_samples_per_second": 76.733,
|
141 |
+
"eval_steps_per_second": 9.592,
|
142 |
+
"step": 5852
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 7.0,
|
146 |
+
"step": 5852,
|
147 |
+
"total_flos": 1.086463725010944e+16,
|
148 |
+
"train_loss": 0.29683067467794466,
|
149 |
+
"train_runtime": 4067.7631,
|
150 |
+
"train_samples_per_second": 32.868,
|
151 |
+
"train_steps_per_second": 4.11
|
152 |
+
}
|
153 |
+
],
|
154 |
+
"logging_steps": 500,
|
155 |
+
"max_steps": 16720,
|
156 |
+
"num_input_tokens_seen": 0,
|
157 |
+
"num_train_epochs": 20,
|
158 |
+
"save_steps": 500,
|
159 |
+
"stateful_callbacks": {
|
160 |
+
"EarlyStoppingCallback": {
|
161 |
+
"args": {
|
162 |
+
"early_stopping_patience": 5,
|
163 |
+
"early_stopping_threshold": 0.0
|
164 |
+
},
|
165 |
+
"attributes": {
|
166 |
+
"early_stopping_patience_counter": 0
|
167 |
+
}
|
168 |
+
},
|
169 |
+
"TrainerControl": {
|
170 |
+
"args": {
|
171 |
+
"should_epoch_stop": false,
|
172 |
+
"should_evaluate": false,
|
173 |
+
"should_log": false,
|
174 |
+
"should_save": true,
|
175 |
+
"should_training_stop": true
|
176 |
+
},
|
177 |
+
"attributes": {}
|
178 |
+
}
|
179 |
+
},
|
180 |
+
"total_flos": 1.086463725010944e+16,
|
181 |
+
"train_batch_size": 8,
|
182 |
+
"trial_name": null,
|
183 |
+
"trial_params": null
|
184 |
+
}
|