jssky commited on
Commit
e2ad41f
1 Parent(s): 7566efc

End of training

Browse files
Files changed (2) hide show
  1. README.md +167 -0
  2. adapter_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,167 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ license: gemma
4
+ base_model: unsloth/gemma-2-2b-it
5
+ tags:
6
+ - axolotl
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: cd45bd18-ae14-49e0-a27b-76db2c9ff747
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/gemma-2-2b-it
23
+ bf16: false
24
+ chat_template: llama3
25
+ dataset_prepared_path: null
26
+ datasets:
27
+ - data_files:
28
+ - 87abe4d80b97880d_train_data.json
29
+ ds_type: json
30
+ format: custom
31
+ path: /workspace/input_data/87abe4d80b97880d_train_data.json
32
+ type:
33
+ field_input: input
34
+ field_instruction: instruction
35
+ field_output: output
36
+ format: '{instruction} {input}'
37
+ no_input_format: '{instruction}'
38
+ system_format: '{system}'
39
+ system_prompt: ''
40
+ debug: null
41
+ deepspeed: null
42
+ devices:
43
+ - 0
44
+ - 1
45
+ - 2
46
+ - 3
47
+ - 4
48
+ - 5
49
+ - 6
50
+ - 7
51
+ early_stopping_patience: null
52
+ eval_max_new_tokens: 128
53
+ eval_table_size: null
54
+ evals_per_epoch: 4
55
+ flash_attention: true
56
+ fp16: true
57
+ fsdp: null
58
+ fsdp_config: null
59
+ gradient_accumulation_steps: 4
60
+ gradient_checkpointing: false
61
+ group_by_length: false
62
+ hub_model_id: jssky/cd45bd18-ae14-49e0-a27b-76db2c9ff747
63
+ hub_repo: null
64
+ hub_strategy: checkpoint
65
+ hub_token: null
66
+ learning_rate: 0.0002
67
+ load_in_4bit: false
68
+ load_in_8bit: false
69
+ local_rank: null
70
+ logging_steps: 1
71
+ lora_alpha: 32
72
+ lora_dropout: 0.05
73
+ lora_fan_in_fan_out: null
74
+ lora_model_dir: null
75
+ lora_r: 16
76
+ lora_target_linear: true
77
+ lr_scheduler: cosine
78
+ max_steps: 10
79
+ micro_batch_size: 1
80
+ mlflow_experiment_name: /tmp/87abe4d80b97880d_train_data.json
81
+ model_type: AutoModelForCausalLM
82
+ num_epochs: 1
83
+ num_gpus: 8
84
+ optimizer: adamw_bnb_8bit
85
+ output_dir: miner_id_24
86
+ pad_to_sequence_len: true
87
+ resume_from_checkpoint: null
88
+ s2_attention: null
89
+ sample_packing: false
90
+ saves_per_epoch: 4
91
+ sequence_len: 4056
92
+ strict: false
93
+ tf32: false
94
+ tokenizer_type: AutoTokenizer
95
+ train_batch_size: 32
96
+ train_on_inputs: false
97
+ trust_remote_code: true
98
+ val_set_size: 0.05
99
+ wandb_entity: null
100
+ wandb_mode: online
101
+ wandb_name: cd45bd18-ae14-49e0-a27b-76db2c9ff747
102
+ wandb_project: Gradients-On-Demand
103
+ wandb_run: your_name
104
+ wandb_runid: cd45bd18-ae14-49e0-a27b-76db2c9ff747
105
+ warmup_steps: 10
106
+ weight_decay: 0.0
107
+ xformers_attention: null
108
+
109
+ ```
110
+
111
+ </details><br>
112
+
113
+ # cd45bd18-ae14-49e0-a27b-76db2c9ff747
114
+
115
+ This model is a fine-tuned version of [unsloth/gemma-2-2b-it](https://huggingface.co/unsloth/gemma-2-2b-it) on the None dataset.
116
+ It achieves the following results on the evaluation set:
117
+ - Loss: 2.0102
118
+
119
+ ## Model description
120
+
121
+ More information needed
122
+
123
+ ## Intended uses & limitations
124
+
125
+ More information needed
126
+
127
+ ## Training and evaluation data
128
+
129
+ More information needed
130
+
131
+ ## Training procedure
132
+
133
+ ### Training hyperparameters
134
+
135
+ The following hyperparameters were used during training:
136
+ - learning_rate: 0.0002
137
+ - train_batch_size: 1
138
+ - eval_batch_size: 1
139
+ - seed: 42
140
+ - distributed_type: multi-GPU
141
+ - num_devices: 8
142
+ - gradient_accumulation_steps: 4
143
+ - total_train_batch_size: 32
144
+ - total_eval_batch_size: 8
145
+ - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
146
+ - lr_scheduler_type: cosine
147
+ - lr_scheduler_warmup_steps: 10
148
+ - training_steps: 10
149
+ - mixed_precision_training: Native AMP
150
+
151
+ ### Training results
152
+
153
+ | Training Loss | Epoch | Step | Validation Loss |
154
+ |:-------------:|:------:|:----:|:---------------:|
155
+ | 2.5602 | 0.0052 | 1 | 2.8289 |
156
+ | 2.7503 | 0.0155 | 3 | 2.6559 |
157
+ | 2.1317 | 0.0311 | 6 | 2.1993 |
158
+ | 2.1536 | 0.0466 | 9 | 2.0102 |
159
+
160
+
161
+ ### Framework versions
162
+
163
+ - PEFT 0.13.2
164
+ - Transformers 4.46.0
165
+ - Pytorch 2.5.0+cu124
166
+ - Datasets 3.0.1
167
+ - Tokenizers 0.20.1
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6802420e455bb7c0a93734e532ff25eed051c4dce431f75a76d41fe57d5c0f8
3
+ size 83197882