base_model: meta-llama/Meta-Llama-3-8B
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: true
load_in_4bit: false
strict: false

datasets:
  - path: translation-dataset-v3-train.hf
    type: alpaca
    train_on_split: train

test_datasets:
  - path: translation-dataset-v3-test.hf
    type: alpaca
    split: train

dataset_prepared_path: ./last_run_prepared
output_dir: ./llama_3_translator
hub_model_id: ahmedsamirio/llama_3_translator_v3


sequence_len: 2048
sample_packing: true
pad_to_sequence_len: true
eval_sample_packing: false

adapter: lora
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project: en_eg_translator
wandb_entity: ahmedsamirio
wandb_name: llama_3_en_eg_translator_v3

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 2
optimizer: paged_adamw_32bit
lr_scheduler: cosine
learning_rate: 2e-5

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 10
evals_per_epoch: 10
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  pad_token: <|end_of_text|>

Visualize in Weights & Biases

Egyptian Arabic Translator Llama-3 8B

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B on the ahmedsamirio/oasst2-9k-translation dataset.

Model description

This model is an attempt to create a small translation model from English to Egyptian Arabic.

Intended uses & limitations

Translating instruction finetuning and text generation datasets

Inference code

python

from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline

tokenizer = AutoTokenizer.from_pretrained("ahmedsamirio/Egyptian-Arabic-Translator-Llama-3-8B")
model = AutoModelForCausalLM.from_pretrained("ahmedsamirio/Egyptian-Arabic-Translator-Llama-3-8B")
pipe = pipeline(task='text-generation', model=model, tokenizer=tokenizer)


en_template = """<|begin_of_text|>Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
Translate the following text to English.

### Input:
{text}

### Response:
"""

ar_template = """<|begin_of_text|>Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
Translate the following text to Arabic.

### Input:
{text}

### Response:
"""

eg_template = """<|begin_of_text|>Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
Translate the following text to Egyptian Arabic.

### Input:
{text}

### Response:
"""

text = """Some habits are known as "keystone habits," and these influence the formation of other habits. \
For example, identifying as the type of person who takes care of their body and is in the habit of exercising regularly, \
can also influence eating better and using credit cards less. In business, \
safety can be a keystone habit that influences other habits that result in greater productivity.[17]"""

ar_text = pipe(ar_template.format(text=text), 
               max_new_tokens=256, 
               do_sample=True, 
               temperature=0.3, 
               top_p=0.5)


eg_text = pipe(eg_template.format(text=ar_text), 
               max_new_tokens=256, 
               do_sample=True, 
               temperature=0.3, 
               top_p=0.5)

print("Original Text:" text)
print("\nArabic Translation:", ar_text)
print("\nEgyptian Arabic Translation:", eg_text)

Training and evaluation data

ahmedsamirio/oasst2-9k-translation

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
0.9661	0.0008	1	1.3816
0.5611	0.1002	123	0.9894
0.6739	0.2004	246	0.8820
0.5168	0.3006	369	0.8229
0.5582	0.4008	492	0.7931
0.552	0.5010	615	0.7814
0.5129	0.6012	738	0.7591
0.5887	0.7014	861	0.7444
0.6359	0.8016	984	0.7293
0.613	0.9018	1107	0.7179
0.5671	1.0020	1230	0.7126
0.4956	1.0847	1353	0.7034
0.5055	1.1849	1476	0.6980
0.4863	1.2851	1599	0.6877
0.4538	1.3853	1722	0.6845
0.4362	1.4855	1845	0.6803
0.4291	1.5857	1968	0.6834
0.6208	1.6859	2091	0.6830
0.582	1.7862	2214	0.6781
0.5001	1.8864	2337	0.6798

Framework versions

PEFT 0.11.1
Transformers 4.42.3
Pytorch 2.1.2+cu118
Datasets 2.19.1
Tokenizers 0.19.1