Update library_name to transformers, add paper and project page links
#1
by
nielsr
HF Staff
- opened
README.md
CHANGED
|
@@ -1,35 +1,37 @@
|
|
| 1 |
---
|
| 2 |
-
|
|
|
|
| 3 |
datasets:
|
| 4 |
-
|
| 5 |
language:
|
| 6 |
-
|
|
|
|
|
|
|
| 7 |
metrics:
|
| 8 |
-
|
| 9 |
-
base_model:
|
| 10 |
-
- meta-llama/Llama-3.3-70B-Instruct
|
| 11 |
pipeline_tag: text-classification
|
| 12 |
-
library_name: adapter-transformers
|
| 13 |
tags:
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
pretty_name: Atlas (LLaMA-3.3-70B) β HTS Classification
|
| 23 |
authors:
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
---
|
| 30 |
|
| 31 |
# Atlas β LLaMA-3.3-70B fine-tuned for Harmonized Tariff Schedule (HTS) classification
|
| 32 |
|
|
|
|
|
|
|
| 33 |
Atlas is a domain-specialized LLaMA-3.3-70B model fine-tuned on U.S. Customs CROSS rulings for Harmonized Tariff Schedule (HTS) code assignment.
|
| 34 |
It targets both **10-digit U.S. HTS (compliance)** and **6-digit HS (globally harmonized)** accuracy.
|
| 35 |
|
|
@@ -41,6 +43,7 @@ Atlas outperforms general-purpose LLMs while remaining deployable/self-hostable.
|
|
| 41 |
- **Model repo:** [flexifyai/atlas-llama3.3-70b-hts-classification](https://huggingface.co/flexifyai/atlas-llama3.3-70b-hts-classification)
|
| 42 |
- **Dataset:** [flexifyai/cross_rulings_hts_dataset_for_tariffs](https://huggingface.co/datasets/flexifyai/cross_rulings_hts_dataset_for_tariffs)
|
| 43 |
- **Demo:** [flexifyai/atlas-llama3_3-70b-hts-demo](https://flexifyai-atlas-llama3-3-70b-hts-demo.hf.space/?__theme=system&deep_link=auHidY8xF00)
|
|
|
|
| 44 |
|
| 45 |
**Example (from the demo):**
|
| 46 |
|
|
@@ -155,4 +158,5 @@ If you find this work useful, please cite our paper:
|
|
| 155 |
archivePrefix={arXiv},
|
| 156 |
primaryClass={cs.AI},
|
| 157 |
url={https://arxiv.org/abs/2509.18400},
|
| 158 |
-
}
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
base_model:
|
| 3 |
+
- meta-llama/Llama-3.3-70B-Instruct
|
| 4 |
datasets:
|
| 5 |
+
- flexifyai/cross_rulings_hts_dataset_for_tariffs
|
| 6 |
language:
|
| 7 |
+
- en
|
| 8 |
+
library_name: transformers
|
| 9 |
+
license: mit
|
| 10 |
metrics:
|
| 11 |
+
- accuracy
|
|
|
|
|
|
|
| 12 |
pipeline_tag: text-classification
|
|
|
|
| 13 |
tags:
|
| 14 |
+
- legal
|
| 15 |
+
- trade
|
| 16 |
+
- htsus
|
| 17 |
+
- semiconductor
|
| 18 |
+
- tariffs
|
| 19 |
+
- hts
|
| 20 |
+
- cross
|
| 21 |
+
- cbp
|
| 22 |
pretty_name: Atlas (LLaMA-3.3-70B) β HTS Classification
|
| 23 |
authors:
|
| 24 |
+
- name: Pritish Yuvraj
|
| 25 |
+
affiliation: Flexify.AI
|
| 26 |
+
homepage: https://www.pritishyuvraj.com/
|
| 27 |
+
- name: Siva Devarakonda
|
| 28 |
+
affiliation: Flexify.AI
|
| 29 |
---
|
| 30 |
|
| 31 |
# Atlas β LLaMA-3.3-70B fine-tuned for Harmonized Tariff Schedule (HTS) classification
|
| 32 |
|
| 33 |
+
This model is presented in the paper [ATLAS: Benchmarking and Adapting LLMs for Global Trade via Harmonized Tariff Code Classification](https://huggingface.co/papers/2509.18400).
|
| 34 |
+
|
| 35 |
Atlas is a domain-specialized LLaMA-3.3-70B model fine-tuned on U.S. Customs CROSS rulings for Harmonized Tariff Schedule (HTS) code assignment.
|
| 36 |
It targets both **10-digit U.S. HTS (compliance)** and **6-digit HS (globally harmonized)** accuracy.
|
| 37 |
|
|
|
|
| 43 |
- **Model repo:** [flexifyai/atlas-llama3.3-70b-hts-classification](https://huggingface.co/flexifyai/atlas-llama3.3-70b-hts-classification)
|
| 44 |
- **Dataset:** [flexifyai/cross_rulings_hts_dataset_for_tariffs](https://huggingface.co/datasets/flexifyai/cross_rulings_hts_dataset_for_tariffs)
|
| 45 |
- **Demo:** [flexifyai/atlas-llama3_3-70b-hts-demo](https://flexifyai-atlas-llama3-3-70b-hts-demo.hf.space/?__theme=system&deep_link=auHidY8xF00)
|
| 46 |
+
- **Project page:** https://tariffpro.flexify.ai/
|
| 47 |
|
| 48 |
**Example (from the demo):**
|
| 49 |
|
|
|
|
| 158 |
archivePrefix={arXiv},
|
| 159 |
primaryClass={cs.AI},
|
| 160 |
url={https://arxiv.org/abs/2509.18400},
|
| 161 |
+
}
|
| 162 |
+
```
|