Card v2: add Zenodo DOI (10.5281/zenodo.19802598) + fix hardware/batch/precision (config drift)

README.md

@@ -21,6 +21,8 @@ pipeline_tag: text-generation
 
 A domain-specific financial reasoning model fine-tuned from [Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) using QLoRA, focused on crypto market analysis, macro reasoning, and multi-step financial logic.
 
+📄 **Paper:** [NPC Fin 32B: A Domain-Specialized Financial Reasoning Model via Multi-GPU QLoRA](https://doi.org/10.5281/zenodo.19802598) (Zenodo, 2026)
+
 ## Model Details
 
 | Parameter | Value |
@@ -32,7 +34,6 @@ A domain-specific financial reasoning model fine-tuned from [Qwen2.5-32B-Instruc
 | LoRA Dropout | 0.05 |
 | Target Modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
 | Training Framework | Unsloth + trl SFTTrainer |
-| Precision | float16 |
 | Max Sequence Length | 4,096 tokens |
 
 ## Training Data
@@ -57,11 +58,19 @@ A domain-specific financial reasoning model fine-tuned from [Qwen2.5-32B-Instruc
 | LR Schedule | Cosine decay |
 | Warmup Ratio | 0.05 |
 | Weight Decay | 0.01 |
-| Batch Size | 4 |
-| Gradient Accumulation | 8 (effective batch size 32) |
+| Per-device Batch Size | 4 |
+| Gradient Accumulation | 8 |
+| Realized Effective Batch | ~384 (4 × 12 GPUs × 8) |
 | Epochs | 3 |
-| Gradient Checkpointing | Enabled |
-| Hardware | NVIDIA A40 48GB (RunPod) |
+| Mixed Precision | bf16 |
+| Distributed Strategy | DeepSpeed ZeRO-3 + full CPU offload |
+| Hardware | 12 × NVIDIA H100 SXM5 80GB (RunPod single multi-GPU node) |
+| Wall Clock | ~72 hours (3 days) |
+| Total Compute | ~864 H100-hours |
+
+
+
+> **Note on batch size:** an earlier version of this card listed the per-device batch (4) and gradient accumulation (8) with an "effective batch 32" annotation inherited from a single-GPU experimental plan. The realized run distributed across 12 H100 GPUs under DeepSpeed ZeRO-3 scaled the effective batch by world size to approximately 384 (4 × 12 × 8). The peak learning rate of 2e-4 was tuned for the planned eff-batch 32, not for the realized 384; see the [paper §4.3 (Config drift: planned vs realized batch size)](https://doi.org/10.5281/zenodo.19802598) for full discussion.
 
 ## Evaluation
 
@@ -120,12 +129,18 @@ print(tokenizer.decode(outputs[0], skip_special_tokens=True))
 
 ## Citation
 
+> Bachu, R. K. (2026). *NPC Fin 32B: A Domain-Specialized Financial Reasoning Model via Multi-GPU QLoRA.* Zenodo. https://doi.org/10.5281/zenodo.19802598
+
 ```bibtex
-@misc{bachu2026npcfin,
-  title={NPC Fin 32B SFT: Domain-Specific Financial Reasoning via QLoRA},
-  author={Ramakrishna Bachu},
-  year={2026},
-  url={https://huggingface.co/ramankrishna10/npc-fin-32b-sft}
+@misc{bachu2026npcfin32b,
+  title     = {NPC Fin 32B: A Domain-Specialized Financial Reasoning Model
+               via Multi-GPU QLoRA},
+  author    = {Bachu, Rama Krishna},
+  year      = {2026},
+  publisher = {Zenodo},
+  doi       = {10.5281/zenodo.19802598},
+  url       = {https://doi.org/10.5281/zenodo.19802598},
+  note      = {Preprint},
 }
 ```