Hugging Face's logo

martimfasantos
/
tinyllama-1.1b-sum-sft-qlora

PEFT
TensorBoard
Safetensors
llama
alignment-handbook
trl
sft
Generated from Trainer
4-bit precision
bitsandbytes
Model card Files Metrics Community
martimfasantos commited on
Commit
3f9078a
•
1 Parent(s): 8122a7a

Model save

Browse files
Files changed (6) hide show
  1. README.md +10 -9
  2. adapter_model.safetensors +1 -1
  3. all_results.json +5 -10
  4. runs/May02_12-16-28_poseidon/events.out.tfevents.1714652205.poseidon.496628.0 +2 -2
  5. train_results.json +5 -5
  6. trainer_state.json +0 -0
README.md CHANGED
@@ -2,13 +2,12 @@
2
  license: apache-2.0
3
  library_name: peft
4
  tags:
5
- - alignment-handbook
6
  - trl
7
  - sft
8
  - generated_from_trainer
9
  base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
10
  datasets:
11
- - martimfasantos/openai-tldr-filtered
12
  model-index:
13
  - name: tinyllama-1.1b-sum-sft-qlora
14
  results: []
@@ -19,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  # tinyllama-1.1b-sum-sft-qlora
21
 
22
- This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T) on the martimfasantos/openai-tldr-filtered dataset.
23
  It achieves the following results on the evaluation set:
24
- - Loss: 2.2356
25
 
26
  ## Model description
27
 
@@ -40,23 +39,25 @@ More information needed
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
- - learning_rate: 6.35e-05
44
  - train_batch_size: 32
45
  - eval_batch_size: 8
46
  - seed: 42
47
  - distributed_type: multi-GPU
48
- - gradient_accumulation_steps: 4
49
- - total_train_batch_size: 128
50
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
51
  - lr_scheduler_type: cosine
52
  - lr_scheduler_warmup_ratio: 0.1
53
- - num_epochs: 1
54
 
55
  ### Training results
56
 
57
  | Training Loss | Epoch | Step | Validation Loss |
58
  |:-------------:|:-----:|:----:|:---------------:|
59
- | 2.2321 | 1.0 | 670 | 2.2356 |
 
 
60
 
61
 
62
  ### Framework versions
 
2
  license: apache-2.0
3
  library_name: peft
4
  tags:
 
5
  - trl
6
  - sft
7
  - generated_from_trainer
8
  base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
9
  datasets:
10
+ - generator
11
  model-index:
12
  - name: tinyllama-1.1b-sum-sft-qlora
13
  results: []
 
18
 
19
  # tinyllama-1.1b-sum-sft-qlora
20
 
21
+ This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T) on the generator dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 2.1637
24
 
25
  ## Model description
26
 
 
39
  ### Training hyperparameters
40
 
41
  The following hyperparameters were used during training:
42
+ - learning_rate: 0.0004
43
  - train_batch_size: 32
44
  - eval_batch_size: 8
45
  - seed: 42
46
  - distributed_type: multi-GPU
47
+ - gradient_accumulation_steps: 2
48
+ - total_train_batch_size: 64
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
  - lr_scheduler_type: cosine
51
  - lr_scheduler_warmup_ratio: 0.1
52
+ - num_epochs: 3
53
 
54
  ### Training results
55
 
56
  | Training Loss | Epoch | Step | Validation Loss |
57
  |:-------------:|:-----:|:----:|:---------------:|
58
+ | 2.1729 | 1.0 | 1340 | 2.1778 |
59
+ | 2.1327 | 2.0 | 2681 | 2.1632 |
60
+ | 2.0989 | 3.0 | 4020 | 2.1637 |
61
 
62
 
63
  ### Framework versions
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c7d7f5577fde7427d431fbd421e7834d5ebdb6f211830165cb775a194bd9e619
3
  size 25272360
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a01995d6e4587c3194bf004a92ba178f15cd3c362591b2794519ccd38f3497cb
3
  size 25272360
all_results.json CHANGED
@@ -1,13 +1,8 @@
1
  {
2
- "epoch": 1.0,
3
- "eval_loss": 2.2356441020965576,
4
- "eval_runtime": 186.9806,
5
- "eval_samples": 6553,
6
- "eval_samples_per_second": 25.789,
7
- "eval_steps_per_second": 3.225,
8
- "train_loss": 2.275535515884855,
9
- "train_runtime": 10859.5951,
10
  "train_samples": 116722,
11
- "train_samples_per_second": 7.898,
12
- "train_steps_per_second": 0.062
13
  }
 
1
  {
2
+ "epoch": 3.0,
3
+ "train_loss": 2.1598364234563725,
4
+ "train_runtime": 32657.7671,
 
 
 
 
 
5
  "train_samples": 116722,
6
+ "train_samples_per_second": 7.879,
7
+ "train_steps_per_second": 0.123
8
  }
runs/May02_12-16-28_poseidon/events.out.tfevents.1714652205.poseidon.496628.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e8d1ebfb8ebeddb638baca56fc89700840e61e0baede73289f97dc3051651ace
3
- size 174733
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3de099420468f8d02ca68a8f8179707aef426ee6939d3e04027817dfe64ed615
3
+ size 176202
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 1.0,
3
- "train_loss": 2.275535515884855,
4
- "train_runtime": 10859.5951,
5
  "train_samples": 116722,
6
- "train_samples_per_second": 7.898,
7
- "train_steps_per_second": 0.062
8
  }
 
1
  {
2
+ "epoch": 3.0,
3
+ "train_loss": 2.1598364234563725,
4
+ "train_runtime": 32657.7671,
5
  "train_samples": 116722,
6
+ "train_samples_per_second": 7.879,
7
+ "train_steps_per_second": 0.123
8
  }
trainer_state.json CHANGED
The diff for this file is too large to render. See raw diff