pico-decoder-medium-1 at 50k steps set to main

README.md

@@ -0,0 +1,52 @@
+---
+license: apache-2.0
+datasets:
+- pico-lm/pretokenized-dolma
+language:
+- en
+metrics:
+- pico-lm/perplexity
+pipeline_tag: text-generation
+---
+
+# Pico Decoder Medium
+
+**pico-decoder-medium** is a 181M parameter model in the `pico-decoder` suite, balancing scale and analyzability. Built with [`pico-train`](https://github.com/pico-lm) and instrumented with [`pico-analyze`](https://github.com/pico-lm), it enables detailed studies of layer-wise learning behavior during language model pretraining.
+
+## 🔧 Model Details
+
+| Field               | Value                              |
+|---------------------|------------------------------------|
+| **Architecture**     | Decoder-only transformer (LLaMA-style) |
+| **Parameters**       | 181M                              |
+| **Layers**           | 12                                |
+| **Hidden Size**      | 768                               |
+| **Feed Forward Size**| 3072                              |
+| **Attention Heads**  | 12                                 |
+| **Key/Value Heads**  | 4                                 |
+
+## 📚 Training
+
+- **Dataset**: [`pretokenized-dolma`](https://github.com/pico-lm)
+- **Training steps**: 200,000
+- **Batch size**: 1024
+- **Sequence length**: 2048
+- **Optimizer**: AdamW
+- **Learning rate schedule**: Linear decay with warmup
+- **Compute**: 16 A100-SXM4-80GB GPUs
+
+## 📈 Evaluation and Analysis
+
+This model supports fine-grained analysis using [pico-analyze](https://github.com/pico-lm). This tool enables researchers to understand how learning unfolds over training, even at very small scales.
+
+We also evaluate perplexity of the model on the [pico-paloma-tinsy](https://huggingface.co/datasets/pico-lm/pretokenized-paloma-tinsy) dataset.
+
+## 📄 Citation
+
+```bibtex
+@software{pico2025,
+    author = {Diehl Martinez, Richard},
+    title = {Pico: A Lightweight Framework for Studying Language Model Learning Dynamics},
+    year = {2025},
+    url = {https://github.com/pico-lm}
+}

config.json

@@ -0,0 +1,22 @@
+{
+  "activation_hidden_dim": 3072,
+  "architectures": [
+    "PicoHF"
+  ],
+  "attention_n_heads": 12,
+  "attention_n_kv_heads": 4,
+  "auto_map": {
+    "AutoConfig": "pico.PicoHFConfig",
+    "AutoModelForCausalLM": "pico.PicoHF"
+  },
+  "batch_size": 1024,
+  "d_model": 768,
+  "max_seq_len": 2048,
+  "model_type": "pico",
+  "n_layers": 12,
+  "norm_eps": 1e-06,
+  "position_emb_theta": 10000.0,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.1",
+  "vocab_size": 50304
+}

eval_results/step_0.json

@@ -0,0 +1 @@
+{"paloma": 59416.7212543554}

eval_results/step_1000.json

@@ -0,0 +1 @@
+{"paloma": 499.5738274564311}

eval_results/step_10000.json

@@ -0,0 +1 @@
+{"paloma": 51.102903796073036}

eval_results/step_11000.json

@@ -0,0 +1 @@
+{"paloma": 48.70718716470207}

eval_results/step_12000.json

@@ -0,0 +1 @@
+{"paloma": 48.04790251246728}

eval_results/step_13000.json

@@ -0,0 +1 @@
+{"paloma": 45.9789450450226}

eval_results/step_14000.json

@@ -0,0 +1 @@
+{"paloma": 45.154312149988236}

eval_results/step_15000.json

@@ -0,0 +1 @@
+{"paloma": 44.130179383032}

eval_results/step_16000.json

@@ -0,0 +1 @@
+{"paloma": 43.383745260105734}

eval_results/step_17000.json

@@ -0,0 +1 @@
+{"paloma": 42.70362300017154}

eval_results/step_18000.json

@@ -0,0 +1 @@
+{"paloma": 42.00629499373951}

eval_results/step_19000.json

@@ -0,0 +1 @@
+{"paloma": 41.85491885225117}

eval_results/step_2000.json

@@ -0,0 +1 @@
+{"paloma": 177.08616091166638}

eval_results/step_20000.json

@@ -0,0 +1 @@
+{"paloma": 40.92828142551595}

eval_results/step_21000.json

@@ -0,0 +1 @@
+{"paloma": 40.3970818064354}

eval_results/step_22000.json

@@ -0,0 +1 @@
+{"paloma": 40.0735236918054}

eval_results/step_23000.json

@@ -0,0 +1 @@
+{"paloma": 39.55490014910283}

eval_results/step_24000.json

@@ -0,0 +1 @@
+{"paloma": 39.20364000381908}

eval_results/step_25000.json

@@ -0,0 +1 @@
+{"paloma": 38.902612380283635}

eval_results/step_26000.json

@@ -0,0 +1 @@
+{"paloma": 38.46148998878559}

eval_results/step_27000.json

@@ -0,0 +1 @@
+{"paloma": 38.17645851178452}

eval_results/step_28000.json

@@ -0,0 +1 @@
+{"paloma": 37.92720969861393}

eval_results/step_29000.json

@@ -0,0 +1 @@
+{"paloma": 37.60208778298275}

eval_results/step_3000.json

@@ -0,0 +1 @@
+{"paloma": 104.02560207121049}

eval_results/step_30000.json

@@ -0,0 +1 @@
+{"paloma": 37.333580768814485}

eval_results/step_31000.json

@@ -0,0 +1 @@
+{"paloma": 37.089247246154095}

eval_results/step_32000.json

@@ -0,0 +1 @@
+{"paloma": 36.82174229962485}

eval_results/step_33000.json

@@ -0,0 +1 @@
+{"paloma": 36.57921791010202}

eval_results/step_34000.json

@@ -0,0 +1 @@
+{"paloma": 36.40737543064543}

eval_results/step_35000.json

@@ -0,0 +1 @@
+{"paloma": 36.21668368017217}

eval_results/step_36000.json

@@ -0,0 +1 @@
+{"paloma": 35.997935054028076}

eval_results/step_37000.json

@@ -0,0 +1 @@
+{"paloma": 35.9598933787296}

eval_results/step_38000.json

@@ -0,0 +1 @@
+{"paloma": 35.7666209674463}

eval_results/step_39000.json

@@ -0,0 +1 @@
+{"paloma": 35.641749255914725}

eval_results/step_4000.json

@@ -0,0 +1 @@
+{"paloma": 85.37458271240939}

eval_results/step_40000.json

@@ -0,0 +1 @@
+{"paloma": 35.34004565797201}

eval_results/step_41000.json

@@ -0,0 +1 @@
+{"paloma": 35.26165773710723}

eval_results/step_42000.json

@@ -0,0 +1 @@
+{"paloma": 35.171348174606884}

eval_results/step_43000.json

@@ -0,0 +1 @@
+{"paloma": 34.918753832391744}

eval_results/step_44000.json

@@ -0,0 +1 @@
+{"paloma": 34.810106504503445}

eval_results/step_45000.json

@@ -0,0 +1 @@
+{"paloma": 34.69993601676064}

eval_results/step_46000.json

@@ -0,0 +1 @@
+{"paloma": 34.59661542430572}

eval_results/step_47000.json

@@ -0,0 +1 @@
+{"paloma": 34.41210062786262}

eval_results/step_48000.json

@@ -0,0 +1 @@
+{"paloma": 34.33602121640581}

eval_results/step_49000.json

@@ -0,0 +1 @@
+{"paloma": 34.21125884014555}

eval_results/step_5000.json

@@ -0,0 +1 @@
+{"paloma": 1283.524473396328}

eval_results/step_50000.json

@@ -0,0 +1 @@
+{"paloma": 34.043001136962545}

eval_results/step_6000.json

@@ -0,0 +1 @@
+{"paloma": 77.90012125104977}