leixa commited on
Commit
92f2712
·
verified ·
1 Parent(s): 506c486

End of training

Browse files
Files changed (2) hide show
  1. README.md +161 -0
  2. adapter_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,161 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ license: apache-2.0
4
+ base_model: TinyLlama/TinyLlama-1.1B-Chat-v0.6
5
+ tags:
6
+ - axolotl
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: 9b0ce33a-0e6b-4448-866c-d786f064ba67
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: TinyLlama/TinyLlama-1.1B-Chat-v0.6
23
+ bf16: auto
24
+ chat_template: llama3
25
+ dataset_prepared_path: null
26
+ datasets:
27
+ - data_files:
28
+ - 652c37adfdba156e_train_data.json
29
+ ds_type: json
30
+ format: custom
31
+ path: /workspace/input_data/652c37adfdba156e_train_data.json
32
+ type:
33
+ field_instruction: prompt
34
+ field_output: chosen
35
+ format: '{instruction}'
36
+ no_input_format: '{instruction}'
37
+ system_format: '{system}'
38
+ system_prompt: ''
39
+ debug: null
40
+ deepspeed: null
41
+ early_stopping_patience: null
42
+ eval_max_new_tokens: 128
43
+ eval_table_size: null
44
+ evals_per_epoch: 4
45
+ flash_attention: true
46
+ fp16: null
47
+ fsdp: null
48
+ fsdp_config: null
49
+ gradient_accumulation_steps: 4
50
+ gradient_checkpointing: true
51
+ gradient_clipping: 1.0
52
+ group_by_length: false
53
+ hub_model_id: leixa/9b0ce33a-0e6b-4448-866c-d786f064ba67
54
+ hub_repo: null
55
+ hub_strategy: checkpoint
56
+ hub_token: null
57
+ learning_rate: 5.0e-05
58
+ load_in_4bit: false
59
+ load_in_8bit: false
60
+ local_rank: 0
61
+ logging_steps: 3
62
+ lora_alpha: 128
63
+ lora_dropout: 0.05
64
+ lora_fan_in_fan_out: null
65
+ lora_model_dir: null
66
+ lora_r: 64
67
+ lora_target_linear: true
68
+ lr_scheduler: cosine
69
+ max_steps: 500
70
+ micro_batch_size: 8
71
+ mlflow_experiment_name: /tmp/652c37adfdba156e_train_data.json
72
+ model_type: AutoModelForCausalLM
73
+ num_epochs: 3
74
+ optimizer: adamw_bnb_8bit
75
+ output_dir: miner_id_24
76
+ pad_to_sequence_len: true
77
+ resume_from_checkpoint: null
78
+ s2_attention: null
79
+ sample_packing: false
80
+ saves_per_epoch: 4
81
+ sequence_len: 1024
82
+ strict: false
83
+ tf32: false
84
+ tokenizer_type: AutoTokenizer
85
+ train_on_inputs: false
86
+ trust_remote_code: true
87
+ val_set_size: 0.05
88
+ wandb_entity: leixa-personal
89
+ wandb_mode: online
90
+ wandb_name: f3658745-2da6-44ad-aeb6-c4275a7f6186
91
+ wandb_project: Gradients-On-Demand
92
+ wandb_run: your_name
93
+ wandb_runid: f3658745-2da6-44ad-aeb6-c4275a7f6186
94
+ warmup_steps: 10
95
+ weight_decay: 0.01
96
+ xformers_attention: null
97
+
98
+ ```
99
+
100
+ </details><br>
101
+
102
+ # 9b0ce33a-0e6b-4448-866c-d786f064ba67
103
+
104
+ This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-Chat-v0.6](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v0.6) on the None dataset.
105
+ It achieves the following results on the evaluation set:
106
+ - Loss: 1.3730
107
+
108
+ ## Model description
109
+
110
+ More information needed
111
+
112
+ ## Intended uses & limitations
113
+
114
+ More information needed
115
+
116
+ ## Training and evaluation data
117
+
118
+ More information needed
119
+
120
+ ## Training procedure
121
+
122
+ ### Training hyperparameters
123
+
124
+ The following hyperparameters were used during training:
125
+ - learning_rate: 5e-05
126
+ - train_batch_size: 8
127
+ - eval_batch_size: 8
128
+ - seed: 42
129
+ - gradient_accumulation_steps: 4
130
+ - total_train_batch_size: 32
131
+ - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
132
+ - lr_scheduler_type: cosine
133
+ - lr_scheduler_warmup_steps: 10
134
+ - training_steps: 288
135
+
136
+ ### Training results
137
+
138
+ | Training Loss | Epoch | Step | Validation Loss |
139
+ |:-------------:|:------:|:----:|:---------------:|
140
+ | No log | 0.0104 | 1 | 1.9109 |
141
+ | 1.6823 | 0.25 | 24 | 1.6583 |
142
+ | 1.5419 | 0.5 | 48 | 1.5523 |
143
+ | 1.5028 | 0.75 | 72 | 1.4936 |
144
+ | 1.4711 | 1.0 | 96 | 1.4529 |
145
+ | 1.4095 | 1.25 | 120 | 1.4250 |
146
+ | 1.4025 | 1.5 | 144 | 1.4069 |
147
+ | 1.3641 | 1.75 | 168 | 1.3919 |
148
+ | 1.3483 | 2.0 | 192 | 1.3826 |
149
+ | 1.3427 | 2.25 | 216 | 1.3779 |
150
+ | 1.3168 | 2.5 | 240 | 1.3749 |
151
+ | 1.3198 | 2.75 | 264 | 1.3732 |
152
+ | 1.3255 | 3.0 | 288 | 1.3730 |
153
+
154
+
155
+ ### Framework versions
156
+
157
+ - PEFT 0.13.2
158
+ - Transformers 4.46.0
159
+ - Pytorch 2.5.0+cu124
160
+ - Datasets 3.0.1
161
+ - Tokenizers 0.20.1
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02752ac69493d8512d7fdf6e1461250dd727e2527d38874e792c9a6c41372188
3
+ size 201962138