Datasets:

Chamaka8
/

serendip-cpt-sinhala

Tasks:

Text Generation

Modalities:

Formats:

Languages:

Size:

Tags:

continual-pretraining

Libraries:

License:

Dataset card Data Studio Files Files and versions

serendip-cpt-sinhala / train_scripts

18.7 kB

1 contributor

History: 4 commits

Chamaka8's picture

Add clean CPT shards, rebuild scripts, and evaluation proof files

6638ccd 7 months ago

load_data.py

940 Bytes
Add training scripts for CPT, SFT-LoRA, DPO and RunPod support 8 months ago
merge_lora.py

616 Bytes
Add training scripts for CPT, SFT-LoRA, DPO and RunPod support 8 months ago
rebuild_clean_shards.py

1.53 kB
Add clean CPT shards, rebuild scripts, and evaluation proof files 7 months ago
rebuild_clean_shards_from_local.py

2.75 kB
Add clean CPT shards, rebuild scripts, and evaluation proof files 7 months ago
requirements.txt

106 Bytes
Add training scripts (CPT baseline + dataset sanity check) 8 months ago
sample_preview.py

415 Bytes
Add training scripts for CPT, SFT-LoRA, DPO and RunPod support 8 months ago
sanity_check_dataset.py

318 Bytes
Add training scripts (CPT baseline + dataset sanity check) 8 months ago
sanity_check_local_jsonl.py

827 Bytes
Add training scripts for CPT, SFT-LoRA, DPO and RunPod support 8 months ago
train_cpt.py

7.33 kB
Add clean CPT shards, rebuild scripts, and evaluation proof files 7 months ago
train_dpo.py

1.39 kB
Add training scripts for CPT, SFT-LoRA, DPO and RunPod support 8 months ago
train_sft_lora.py

2.45 kB
Add training scripts for CPT, SFT-LoRA, DPO and RunPod support 8 months ago