dimasik1987 commited on
Commit
e52add6
·
verified ·
1 Parent(s): 0c270f5

End of training

Browse files
Files changed (2) hide show
  1. README.md +159 -0
  2. adapter_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,159 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ base_model: peft-internal-testing/tiny-dummy-qwen2
4
+ tags:
5
+ - axolotl
6
+ - generated_from_trainer
7
+ model-index:
8
+ - name: 8e058ffa-6204-4515-8a3f-61d909a20da4
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
16
+ <details><summary>See axolotl config</summary>
17
+
18
+ axolotl version: `0.4.1`
19
+ ```yaml
20
+ adapter: lora
21
+ base_model: peft-internal-testing/tiny-dummy-qwen2
22
+ bf16: auto
23
+ chat_template: llama3
24
+ dataset_prepared_path: null
25
+ datasets:
26
+ - data_files:
27
+ - 8d8edba61bbb4a8e_train_data.json
28
+ ds_type: json
29
+ format: custom
30
+ path: /workspace/input_data/8d8edba61bbb4a8e_train_data.json
31
+ type:
32
+ field_instruction: error_sentence
33
+ field_output: correct_sentence
34
+ format: '{instruction}'
35
+ no_input_format: '{instruction}'
36
+ system_format: '{system}'
37
+ system_prompt: ''
38
+ debug: null
39
+ deepspeed: null
40
+ device: cuda
41
+ early_stopping_patience: 1
42
+ eval_max_new_tokens: 128
43
+ eval_steps: 5
44
+ eval_table_size: null
45
+ evals_per_epoch: null
46
+ flash_attention: false
47
+ fp16: null
48
+ gradient_accumulation_steps: 4
49
+ gradient_checkpointing: true
50
+ group_by_length: true
51
+ hub_model_id: dimasik1987/8e058ffa-6204-4515-8a3f-61d909a20da4
52
+ hub_repo: null
53
+ hub_strategy: checkpoint
54
+ hub_token: null
55
+ learning_rate: 0.0002
56
+ load_in_4bit: false
57
+ load_in_8bit: false
58
+ local_rank: null
59
+ logging_steps: 3
60
+ lora_alpha: 32
61
+ lora_dropout: 0.05
62
+ lora_fan_in_fan_out: null
63
+ lora_model_dir: null
64
+ lora_r: 16
65
+ lora_target_linear: true
66
+ lr_scheduler: cosine
67
+ max_memory:
68
+ 0: 79GiB
69
+ max_steps: 30
70
+ micro_batch_size: 4
71
+ mlflow_experiment_name: /tmp/8d8edba61bbb4a8e_train_data.json
72
+ model_type: AutoModelForCausalLM
73
+ num_epochs: 1
74
+ optim_args:
75
+ adam_beta1: 0.9
76
+ adam_beta2: 0.95
77
+ adam_epsilon: 1e-5
78
+ optimizer: adamw_torch
79
+ output_dir: miner_id_24
80
+ pad_to_sequence_len: true
81
+ resume_from_checkpoint: null
82
+ s2_attention: null
83
+ sample_packing: false
84
+ save_steps: 10
85
+ sequence_len: 1024
86
+ strict: false
87
+ tf32: false
88
+ tokenizer_type: AutoTokenizer
89
+ train_on_inputs: true
90
+ trust_remote_code: true
91
+ val_set_size: 0.05
92
+ wandb_entity: null
93
+ wandb_mode: online
94
+ wandb_name: fcf5881b-2457-4721-807e-7e3e322c193a
95
+ wandb_project: Gradients-On-Demand
96
+ wandb_run: your_name
97
+ wandb_runid: fcf5881b-2457-4721-807e-7e3e322c193a
98
+ warmup_steps: 5
99
+ weight_decay: 0.001
100
+ xformers_attention: true
101
+
102
+ ```
103
+
104
+ </details><br>
105
+
106
+ # 8e058ffa-6204-4515-8a3f-61d909a20da4
107
+
108
+ This model is a fine-tuned version of [peft-internal-testing/tiny-dummy-qwen2](https://huggingface.co/peft-internal-testing/tiny-dummy-qwen2) on the None dataset.
109
+ It achieves the following results on the evaluation set:
110
+ - Loss: 11.9323
111
+
112
+ ## Model description
113
+
114
+ More information needed
115
+
116
+ ## Intended uses & limitations
117
+
118
+ More information needed
119
+
120
+ ## Training and evaluation data
121
+
122
+ More information needed
123
+
124
+ ## Training procedure
125
+
126
+ ### Training hyperparameters
127
+
128
+ The following hyperparameters were used during training:
129
+ - learning_rate: 0.0002
130
+ - train_batch_size: 4
131
+ - eval_batch_size: 4
132
+ - seed: 42
133
+ - gradient_accumulation_steps: 4
134
+ - total_train_batch_size: 16
135
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
136
+ - lr_scheduler_type: cosine
137
+ - lr_scheduler_warmup_steps: 5
138
+ - training_steps: 30
139
+
140
+ ### Training results
141
+
142
+ | Training Loss | Epoch | Step | Validation Loss |
143
+ |:-------------:|:------:|:----:|:---------------:|
144
+ | No log | 0.0009 | 1 | 11.9326 |
145
+ | 11.9345 | 0.0043 | 5 | 11.9325 |
146
+ | 11.9321 | 0.0085 | 10 | 11.9325 |
147
+ | 11.9337 | 0.0128 | 15 | 11.9324 |
148
+ | 11.9335 | 0.0171 | 20 | 11.9323 |
149
+ | 11.9322 | 0.0213 | 25 | 11.9322 |
150
+ | 11.9314 | 0.0256 | 30 | 11.9323 |
151
+
152
+
153
+ ### Framework versions
154
+
155
+ - PEFT 0.13.2
156
+ - Transformers 4.46.0
157
+ - Pytorch 2.5.0+cu124
158
+ - Datasets 3.0.1
159
+ - Tokenizers 0.20.1
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f05dc9d63606a0c5ce78433f5ecf6371a1b98c0534264639a70d6fbc0532efd1
3
+ size 32642