Training in progress, step 1000
Browse files- .gitattributes +1 -0
- README.md +38 -38
- adapter_config.json +61 -0
- adapter_model.safetensors +3 -0
- chat_template.jinja +49 -0
- runs/Jun28_07-47-04_8fa2e96bced5/events.out.tfevents.1751096841.8fa2e96bced5.295305.0 +3 -0
- special_tokens_map.json +36 -0
- tokenizer.json +3 -0
- tokenizer_config.json +0 -0
- training_args.bin +2 -2
.gitattributes
CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
@@ -1,58 +1,58 @@
|
|
1 |
---
|
2 |
-
library_name: transformers
|
3 |
-
license: gemma
|
4 |
base_model: google/gemma-3n-E4B-it
|
|
|
|
|
5 |
tags:
|
6 |
- generated_from_trainer
|
7 |
-
|
8 |
-
-
|
9 |
-
|
10 |
---
|
11 |
|
12 |
-
|
13 |
-
should probably proofread and complete it, then remove this comment. -->
|
14 |
-
|
15 |
-
# gemma-3n-E4B-transcribe-zh-tw-1
|
16 |
-
|
17 |
-
This model is a fine-tuned version of [google/gemma-3n-E4B-it](https://huggingface.co/google/gemma-3n-E4B-it) on an unknown dataset.
|
18 |
|
19 |
-
|
|
|
20 |
|
21 |
-
|
22 |
|
23 |
-
|
|
|
24 |
|
25 |
-
|
|
|
|
|
|
|
|
|
26 |
|
27 |
-
## Training
|
28 |
|
29 |
-
|
30 |
|
31 |
-
## Training procedure
|
32 |
|
33 |
-
|
34 |
|
35 |
-
|
36 |
-
- learning_rate: 1e-05
|
37 |
-
- train_batch_size: 4
|
38 |
-
- eval_batch_size: 4
|
39 |
-
- seed: 42
|
40 |
-
- gradient_accumulation_steps: 2
|
41 |
-
- total_train_batch_size: 8
|
42 |
-
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
43 |
-
- lr_scheduler_type: linear
|
44 |
-
- num_epochs: 3
|
45 |
|
46 |
-
|
|
|
|
|
|
|
|
|
47 |
|
48 |
-
|
49 |
-
|:-------------:|:-----:|:----:|:---------------:|:------:|:---:|
|
50 |
-
| No log | 0 | 0 | 8.2124 | 2.2923 | 1.0 |
|
51 |
|
52 |
|
53 |
-
### Framework versions
|
54 |
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
|
|
|
|
2 |
base_model: google/gemma-3n-E4B-it
|
3 |
+
library_name: transformers
|
4 |
+
model_name: gemma-3n-E4B-transcribe-zh-tw-1
|
5 |
tags:
|
6 |
- generated_from_trainer
|
7 |
+
- trl
|
8 |
+
- sft
|
9 |
+
licence: license
|
10 |
---
|
11 |
|
12 |
+
# Model Card for gemma-3n-E4B-transcribe-zh-tw-1
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
+
This model is a fine-tuned version of [google/gemma-3n-E4B-it](https://huggingface.co/google/gemma-3n-E4B-it).
|
15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
16 |
|
17 |
+
## Quick start
|
18 |
|
19 |
+
```python
|
20 |
+
from transformers import pipeline
|
21 |
|
22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
23 |
+
generator = pipeline("text-generation", model="JacobLinCool/gemma-3n-E4B-transcribe-zh-tw-1", device="cuda")
|
24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
25 |
+
print(output["generated_text"])
|
26 |
+
```
|
27 |
|
28 |
+
## Training procedure
|
29 |
|
30 |
+
|
31 |
|
|
|
32 |
|
33 |
+
This model was trained with SFT.
|
34 |
|
35 |
+
### Framework versions
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
|
37 |
+
- TRL: 0.19.0
|
38 |
+
- Transformers: 4.53.0
|
39 |
+
- Pytorch: 2.8.0.dev20250319+cu128
|
40 |
+
- Datasets: 3.6.0
|
41 |
+
- Tokenizers: 0.21.2
|
42 |
|
43 |
+
## Citations
|
|
|
|
|
44 |
|
45 |
|
|
|
46 |
|
47 |
+
Cite TRL as:
|
48 |
+
|
49 |
+
```bibtex
|
50 |
+
@misc{vonwerra2022trl,
|
51 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
52 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
|
53 |
+
year = 2020,
|
54 |
+
journal = {GitHub repository},
|
55 |
+
publisher = {GitHub},
|
56 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
57 |
+
}
|
58 |
+
```
|
adapter_config.json
ADDED
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "google/gemma-3n-E4B-it",
|
5 |
+
"bias": "none",
|
6 |
+
"corda_config": null,
|
7 |
+
"eva_config": null,
|
8 |
+
"exclude_modules": null,
|
9 |
+
"fan_in_fan_out": false,
|
10 |
+
"inference_mode": true,
|
11 |
+
"init_lora_weights": true,
|
12 |
+
"layer_replication": null,
|
13 |
+
"layers_pattern": null,
|
14 |
+
"layers_to_transform": null,
|
15 |
+
"loftq_config": {},
|
16 |
+
"lora_alpha": 1,
|
17 |
+
"lora_bias": false,
|
18 |
+
"lora_dropout": 0.0,
|
19 |
+
"megatron_config": null,
|
20 |
+
"megatron_core": "megatron.core",
|
21 |
+
"modules_to_save": null,
|
22 |
+
"peft_type": "LORA",
|
23 |
+
"r": 32,
|
24 |
+
"rank_pattern": {},
|
25 |
+
"revision": null,
|
26 |
+
"target_modules": [
|
27 |
+
"up_proj",
|
28 |
+
"altup_unembed_projections.2",
|
29 |
+
"per_layer_projection",
|
30 |
+
"linear_end",
|
31 |
+
"linear_start",
|
32 |
+
"ffw_layer_1",
|
33 |
+
"per_layer_model_projection",
|
34 |
+
"correction_coefs",
|
35 |
+
"pos_proj",
|
36 |
+
"altup_unembed_projections.1",
|
37 |
+
"altup_projections.0",
|
38 |
+
"down_proj",
|
39 |
+
"v_proj",
|
40 |
+
"modality_router",
|
41 |
+
"per_layer_input_gate",
|
42 |
+
"altup_projections.2",
|
43 |
+
"ffw_layer_2",
|
44 |
+
"post",
|
45 |
+
"gate_proj",
|
46 |
+
"prediction_coefs",
|
47 |
+
"o_proj",
|
48 |
+
"q_proj",
|
49 |
+
"linear_right",
|
50 |
+
"altup_unembed_projections.0",
|
51 |
+
"linear_left",
|
52 |
+
"embedding_projection",
|
53 |
+
"input_proj_linear",
|
54 |
+
"k_proj",
|
55 |
+
"altup_projections.1"
|
56 |
+
],
|
57 |
+
"task_type": "CAUSAL_LM",
|
58 |
+
"trainable_token_indices": null,
|
59 |
+
"use_dora": false,
|
60 |
+
"use_rslora": true
|
61 |
+
}
|
adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d82f6b427fab5c6277e902b680eddec235d3401cca47349c7af06c71f4d12d6f
|
3 |
+
size 444884256
|
chat_template.jinja
ADDED
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{{ bos_token }}
|
2 |
+
{%- if messages[0]['role'] == 'system' -%}
|
3 |
+
{%- if messages[0]['content'] is string -%}
|
4 |
+
{%- set first_user_prefix = messages[0]['content'] + '
|
5 |
+
|
6 |
+
' -%}
|
7 |
+
{%- else -%}
|
8 |
+
{%- set first_user_prefix = messages[0]['content'][0]['text'] + '
|
9 |
+
|
10 |
+
' -%}
|
11 |
+
{%- endif -%}
|
12 |
+
{%- set loop_messages = messages[1:] -%}
|
13 |
+
{%- else -%}
|
14 |
+
{%- set first_user_prefix = "" -%}
|
15 |
+
{%- set loop_messages = messages -%}
|
16 |
+
{%- endif -%}
|
17 |
+
{%- for message in loop_messages -%}
|
18 |
+
{%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
|
19 |
+
{{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
|
20 |
+
{%- endif -%}
|
21 |
+
{%- if (message['role'] == 'assistant') -%}
|
22 |
+
{%- set role = "model" -%}
|
23 |
+
{%- else -%}
|
24 |
+
{%- set role = message['role'] -%}
|
25 |
+
{%- endif -%}
|
26 |
+
{{ '<start_of_turn>' + role + '
|
27 |
+
' + (first_user_prefix if loop.first else "") }}
|
28 |
+
{%- if message['content'] is string -%}
|
29 |
+
{{ message['content'] | trim }}
|
30 |
+
{%- elif message['content'] is iterable -%}
|
31 |
+
{%- for item in message['content'] -%}
|
32 |
+
{%- if item['type'] == 'audio' -%}
|
33 |
+
{{ '<audio_soft_token>' }}
|
34 |
+
{%- elif item['type'] == 'image' -%}
|
35 |
+
{{ '<image_soft_token>' }}
|
36 |
+
{%- elif item['type'] == 'text' -%}
|
37 |
+
{{ item['text'] | trim }}
|
38 |
+
{%- endif -%}
|
39 |
+
{%- endfor -%}
|
40 |
+
{%- else -%}
|
41 |
+
{{ raise_exception("Invalid content type") }}
|
42 |
+
{%- endif -%}
|
43 |
+
{{ '<end_of_turn>
|
44 |
+
' }}
|
45 |
+
{%- endfor -%}
|
46 |
+
{%- if add_generation_prompt -%}
|
47 |
+
{{'<start_of_turn>model
|
48 |
+
'}}
|
49 |
+
{%- endif -%}
|
runs/Jun28_07-47-04_8fa2e96bced5/events.out.tfevents.1751096841.8fa2e96bced5.295305.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4c250c26e26e1185ce759f2a46eb9b762892c339db4d1b0da4d7857b2b00a7f6
|
3 |
+
size 338811
|
special_tokens_map.json
ADDED
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"audio_token": "<audio_soft_token>",
|
3 |
+
"boa_token": "<start_of_audio>",
|
4 |
+
"boi_token": "<start_of_image>",
|
5 |
+
"bos_token": {
|
6 |
+
"content": "<bos>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": false,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"eoa_token": "<end_of_audio>",
|
13 |
+
"eoi_token": "<end_of_image>",
|
14 |
+
"eos_token": {
|
15 |
+
"content": "<eos>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": false,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false
|
20 |
+
},
|
21 |
+
"image_token": "<image_soft_token>",
|
22 |
+
"pad_token": {
|
23 |
+
"content": "<pad>",
|
24 |
+
"lstrip": false,
|
25 |
+
"normalized": false,
|
26 |
+
"rstrip": false,
|
27 |
+
"single_word": false
|
28 |
+
},
|
29 |
+
"unk_token": {
|
30 |
+
"content": "<unk>",
|
31 |
+
"lstrip": false,
|
32 |
+
"normalized": false,
|
33 |
+
"rstrip": false,
|
34 |
+
"single_word": false
|
35 |
+
}
|
36 |
+
}
|
tokenizer.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ed504d44c6898db6d71f4abf77cf8e12a9ee3ac90bb1aec0e1edc012c4251b23
|
3 |
+
size 33442708
|
tokenizer_config.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7a2ce00daeadd7c0c8033a71070edcf6b1dd57db3255aae04d041659b28a4f34
|
3 |
+
size 6289
|