This repo contains the continent 1b step2k model exported from the 2k checkpoint for the metadata localization project. It was trained from scratch on the project corpus, using the Llama 3.2 tokenizer and vocabulary.

Variant Metadata

Stage: pretrain
Family: metadata_ablation
Size: 1b
Metadata condition: with_metadata
Checkpoint export: 2k
Base model lineage: Trained from scratch; tokenizer/vocabulary from meta-llama/Llama-3.2-1B

Weights & Biases Provenance

Run name: 23/03/2026_08:58:03_combined_only_continent_with_metadata_1b
Internal run URL: https://wandb.ai/iamshnoo/nanotron/runs/6baqnhow
Note: the Weights & Biases workspace is private; public readers should use the summarized metrics and configuration below.
State: finished
Runtime: 57h 18m 26s

Run Summary

KPI/train_lm_loss: 2.1429
KPI/train_perplexity: 8.5243
KPI/val_loss: 2.1753
KPI/val_perplexity: 8.8048
KPI/consumed_tokens/train: 41,943,040,000
_step: 10,000

Training Configuration

train_steps: 10,000
sequence_length: 2,048
micro_batch_size: 8
batch_accumulation_per_replica: 64
learning_rate: 0.003
min_decay_lr: 0.0003
checkpoint_interval: 1,000

Training Curves

Static plots below were exported from the private Weights & Biases run and embedded here for public access.

Train Loss

Validation Perplexity

Throughput

Project Context

This model is part of the metadata localization release. Related checkpoints and variants are grouped in the public Hugging Face collection Metadata Conditioned LLMs.

Training data source: News on the Web (NOW) Corpus
Project repository: https://github.com/iamshnoo/metadata_localization
Paper: https://arxiv.org/abs/2601.15236

Last synced: 2026-04-02 14:42:51 UTC

iamshnoo/combined_only_continent_with_metadata_1b_step2k

combined_only_continent_with_metadata_1b_step2k

Summary