dada22231 commited on
Commit
4a015b2
1 Parent(s): d46f8df

End of training

Browse files
Files changed (2) hide show
  1. README.md +174 -0
  2. adapter_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,174 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ license: apache-2.0
4
+ base_model: unsloth/tinyllama-chat
5
+ tags:
6
+ - axolotl
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: cdf861c6-72a4-41b4-ac3f-a9bb00013b19
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ [<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)
17
+ <details><summary>See axolotl config</summary>
18
+
19
+ axolotl version: `0.4.1`
20
+ ```yaml
21
+ adapter: lora
22
+ base_model: unsloth/tinyllama-chat
23
+ bf16: auto
24
+ chat_template: llama3
25
+ cosine_min_lr_ratio: 0.1
26
+ data_processes: 4
27
+ dataset_prepared_path: null
28
+ datasets:
29
+ - data_files:
30
+ - 11bfdd43471b04af_train_data.json
31
+ ds_type: json
32
+ format: custom
33
+ num_proc: 4
34
+ path: /workspace/input_data/11bfdd43471b04af_train_data.json
35
+ streaming: true
36
+ type:
37
+ field_instruction: question
38
+ field_output: response
39
+ format: '{instruction}'
40
+ no_input_format: '{instruction}'
41
+ system_format: '{system}'
42
+ system_prompt: ''
43
+ debug: null
44
+ deepspeed: null
45
+ device_map: balanced
46
+ do_eval: true
47
+ early_stopping_patience: 1
48
+ eval_batch_size: 1
49
+ eval_sample_packing: false
50
+ eval_steps: 25
51
+ evaluation_strategy: steps
52
+ flash_attention: false
53
+ fp16: null
54
+ fsdp: null
55
+ fsdp_config: null
56
+ gradient_accumulation_steps: 16
57
+ gradient_checkpointing: true
58
+ group_by_length: true
59
+ hub_model_id: dada22231/cdf861c6-72a4-41b4-ac3f-a9bb00013b19
60
+ hub_strategy: checkpoint
61
+ hub_token: null
62
+ learning_rate: 0.0001
63
+ load_in_4bit: false
64
+ load_in_8bit: false
65
+ local_rank: null
66
+ logging_steps: 1
67
+ lora_alpha: 64
68
+ lora_dropout: 0.05
69
+ lora_fan_in_fan_out: null
70
+ lora_model_dir: null
71
+ lora_r: 32
72
+ lora_target_linear: true
73
+ lora_target_modules:
74
+ - q_proj
75
+ - v_proj
76
+ lr_scheduler: cosine
77
+ max_grad_norm: 1.0
78
+ max_memory:
79
+ 0: 75GB
80
+ 1: 75GB
81
+ 2: 75GB
82
+ 3: 75GB
83
+ max_steps: 50
84
+ micro_batch_size: 2
85
+ mixed_precision: bf16
86
+ mlflow_experiment_name: /tmp/11bfdd43471b04af_train_data.json
87
+ model_type: AutoModelForCausalLM
88
+ num_epochs: 3
89
+ optim_args:
90
+ adam_beta1: 0.9
91
+ adam_beta2: 0.95
92
+ adam_epsilon: 1e-5
93
+ optimizer: adamw_torch
94
+ output_dir: miner_id_24
95
+ pad_to_sequence_len: true
96
+ resume_from_checkpoint: null
97
+ s2_attention: null
98
+ sample_packing: false
99
+ save_steps: 25
100
+ save_strategy: steps
101
+ sequence_len: 2048
102
+ strict: false
103
+ tf32: false
104
+ tokenizer_type: AutoTokenizer
105
+ torch_compile: false
106
+ train_on_inputs: false
107
+ trust_remote_code: true
108
+ val_set_size: 50
109
+ wandb_entity: null
110
+ wandb_mode: online
111
+ wandb_name: cdf861c6-72a4-41b4-ac3f-a9bb00013b19
112
+ wandb_project: Public_TuningSN
113
+ wandb_runid: cdf861c6-72a4-41b4-ac3f-a9bb00013b19
114
+ warmup_ratio: 0.04
115
+ weight_decay: 0.01
116
+ xformers_attention: null
117
+
118
+ ```
119
+
120
+ </details><br>
121
+
122
+ # cdf861c6-72a4-41b4-ac3f-a9bb00013b19
123
+
124
+ This model is a fine-tuned version of [unsloth/tinyllama-chat](https://huggingface.co/unsloth/tinyllama-chat) on the None dataset.
125
+ It achieves the following results on the evaluation set:
126
+ - Loss: nan
127
+
128
+ ## Model description
129
+
130
+ More information needed
131
+
132
+ ## Intended uses & limitations
133
+
134
+ More information needed
135
+
136
+ ## Training and evaluation data
137
+
138
+ More information needed
139
+
140
+ ## Training procedure
141
+
142
+ ### Training hyperparameters
143
+
144
+ The following hyperparameters were used during training:
145
+ - learning_rate: 0.0001
146
+ - train_batch_size: 2
147
+ - eval_batch_size: 1
148
+ - seed: 42
149
+ - distributed_type: multi-GPU
150
+ - num_devices: 4
151
+ - gradient_accumulation_steps: 16
152
+ - total_train_batch_size: 128
153
+ - total_eval_batch_size: 4
154
+ - 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
155
+ - lr_scheduler_type: cosine
156
+ - lr_scheduler_warmup_steps: 2
157
+ - training_steps: 50
158
+
159
+ ### Training results
160
+
161
+ | Training Loss | Epoch | Step | Validation Loss |
162
+ |:-------------:|:------:|:----:|:---------------:|
163
+ | 332.6675 | 0.0012 | 1 | nan |
164
+ | 0.0 | 0.0295 | 25 | nan |
165
+ | 0.0 | 0.0590 | 50 | nan |
166
+
167
+
168
+ ### Framework versions
169
+
170
+ - PEFT 0.13.2
171
+ - Transformers 4.46.0
172
+ - Pytorch 2.5.0+cu124
173
+ - Datasets 3.0.1
174
+ - Tokenizers 0.20.1
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb77c099d494f2cc1e4c03964f43740d07abc45295f7553fb5b587b64eaea4ec
3
+ size 101036698