w-ahmad commited on
Commit
f325e5d
·
verified ·
1 Parent(s): e56395a

Upload entire model folder

Browse files
out/situglu_run/README.md ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ tags:
4
+ - generated_from_trainer
5
+ model-index:
6
+ - name: tiny-llama-baseline-situglu
7
+ results: []
8
+ ---
9
+
10
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
11
+ should probably proofread and complete it, then remove this comment. -->
12
+
13
+ # tiny-llama-baseline-situglu
14
+
15
+ This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
16
+ It achieves the following results on the evaluation set:
17
+ - Loss: 1.3839
18
+
19
+ ## Model description
20
+
21
+ More information needed
22
+
23
+ ## Intended uses & limitations
24
+
25
+ More information needed
26
+
27
+ ## Training and evaluation data
28
+
29
+ More information needed
30
+
31
+ ## Training procedure
32
+
33
+ ### Training hyperparameters
34
+
35
+ The following hyperparameters were used during training:
36
+ - learning_rate: 0.0003
37
+ - train_batch_size: 32
38
+ - eval_batch_size: 32
39
+ - seed: 42
40
+ - gradient_accumulation_steps: 2
41
+ - total_train_batch_size: 64
42
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
43
+ - lr_scheduler_type: constant
44
+ - training_steps: 10000
45
+
46
+ ### Training results
47
+
48
+ | Training Loss | Epoch | Step | Validation Loss |
49
+ |:-------------:|:------:|:-----:|:---------------:|
50
+ | 5.8313 | 0.0159 | 250 | 2.8854 |
51
+ | 4.6325 | 0.0319 | 500 | 2.3345 |
52
+ | 4.1802 | 0.0478 | 750 | 2.0870 |
53
+ | 3.8631 | 0.0637 | 1000 | 1.9404 |
54
+ | 3.7073 | 0.0796 | 1250 | 1.8444 |
55
+ | 3.5027 | 0.0956 | 1500 | 1.7725 |
56
+ | 3.4414 | 0.1115 | 1750 | 1.7238 |
57
+ | 3.3564 | 0.1274 | 2000 | 1.6845 |
58
+ | 3.3003 | 0.1433 | 2250 | 1.6524 |
59
+ | 3.1822 | 0.1593 | 2500 | 1.6235 |
60
+ | 3.1618 | 0.1752 | 2750 | 1.5997 |
61
+ | 3.1667 | 0.1911 | 3000 | 1.5814 |
62
+ | 3.1793 | 0.2071 | 3250 | 1.5652 |
63
+ | 3.1133 | 0.2230 | 3500 | 1.5505 |
64
+ | 3.0683 | 0.2389 | 3750 | 1.5347 |
65
+ | 3.0116 | 0.2548 | 4000 | 1.5229 |
66
+ | 3.0286 | 0.2708 | 4250 | 1.5121 |
67
+ | 2.9835 | 0.2867 | 4500 | 1.5008 |
68
+ | 2.9524 | 0.3026 | 4750 | 1.4927 |
69
+ | 2.9365 | 0.3185 | 5000 | 1.4838 |
70
+ | 2.9148 | 0.3345 | 5250 | 1.4760 |
71
+ | 2.9646 | 0.3504 | 5500 | 1.4690 |
72
+ | 2.9633 | 0.3663 | 5750 | 1.4613 |
73
+ | 2.9265 | 0.3823 | 6000 | 1.4567 |
74
+ | 2.9072 | 0.3982 | 6250 | 1.4485 |
75
+ | 2.8895 | 0.4141 | 6500 | 1.4441 |
76
+ | 2.8846 | 0.4300 | 6750 | 1.4385 |
77
+ | 2.8704 | 0.4460 | 7000 | 1.4334 |
78
+ | 2.8507 | 0.4619 | 7250 | 1.4272 |
79
+ | 2.8462 | 0.4778 | 7500 | 1.4215 |
80
+ | 2.8561 | 0.4937 | 7750 | 1.4190 |
81
+ | 2.7884 | 0.5097 | 8000 | 1.4141 |
82
+ | 2.8249 | 0.5256 | 8250 | 1.4105 |
83
+ | 2.8453 | 0.5415 | 8500 | 1.4060 |
84
+ | 2.7967 | 0.5574 | 8750 | 1.4015 |
85
+ | 2.8201 | 0.5734 | 9000 | 1.3981 |
86
+ | 2.7965 | 0.5893 | 9250 | 1.3942 |
87
+ | 2.7560 | 0.6052 | 9500 | 1.3919 |
88
+ | 2.7896 | 0.6212 | 9750 | 1.3868 |
89
+ | 2.7553 | 0.6371 | 10000 | 1.3839 |
90
+
91
+
92
+ ### Framework versions
93
+
94
+ - Transformers 5.15.0.dev0
95
+ - Pytorch 2.6.0+cu124
96
+ - Datasets 5.0.1
97
+ - Tokenizers 0.22.2
out/situglu_run/chat_template.jinja ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {% for message in messages %}
2
+ {% if message['role'] == 'user' %}
3
+ {{ '<|user|>
4
+ ' + message['content'] + eos_token }}
5
+ {% elif message['role'] == 'system' %}
6
+ {{ '<|system|>
7
+ ' + message['content'] + eos_token }}
8
+ {% elif message['role'] == 'assistant' %}
9
+ {{ '<|assistant|>
10
+ ' + message['content'] + eos_token }}
11
+ {% endif %}
12
+ {% if loop.last and add_generation_prompt %}
13
+ {{ '<|assistant|>' }}
14
+ {% endif %}
15
+ {% endfor %}