Super-squash branch 'main' using huggingface_hub

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,25 @@
+# AGILLM-3.5
+
+698M-param decoder LM (d=1024, 24 layers, 16 heads, rank=128, expansion 2.0×), DeepSeek-V3.2
+tokenizer, AR + SAT (speculative) heads, trained with **DiffusionBlocks** (block-wise EDM
+denoising, 8 blocks). Forked from the AGILLM-3 ~step-51081 base.
+
+## Checkpoints
+
+### `distributed/` — the live distributed model
+`master_round244.pt` — trained **block-disaggregated across 4 Hetzner nodes** (GETH +
+MCP + PRIME + COMMUNIST-WEB). GETH coordinates and trains blocks 0,1,3,5,7 locally; MCP/PRIME/
+COMMUNIST-WEB train blocks 2/4/6 over the private network. Each round exports block slices,
+trains them independently, and merges them back. **739 merged block-updates** at snapshot.
+Single full file per snapshot (each round is a block merge, not a delta).
+
+### `single_node/` — single-node dblock lineage (full + delta)
+- `pretrain_step00053908.pt` — full checkpoint (7.3 GB)
+- `pretrain_delta_step00053702.pt` (+ `.sha256`) — delta checkpoint (2.8 GB)
+
+Checkpoint dict keys: `core` (backbone), `ar`, `sat` (heads), `cfg`, embedded
+`tokenizer_json`, plus `disagg_updates` (merge provenance) on the distributed master.
+
+## Inference
+Load with the AGILLM nB300 code (`infer --mode ar|sat`); the tokenizer round-trips from the
+embedded `tokenizer_json`.

distributed/master_latest.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fde4b297b70b8aa6ed240268214843a9ff6f2767ca1f3c9a8b68723e8a5563e
+size 6569221865

single_node/pretrain_delta_step00053702.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32a9390cfaeb1908eef98c75ca49babd8eda4eba2116c233f9023dd82a077eca
+size 2793593611

single_node/pretrain_delta_step00053702.sha256

@@ -0,0 +1 @@
+32a9390cfaeb1908eef98c75ca49babd8eda4eba2116c233f9023dd82a077eca  pretrain_delta_step00053702.pt

single_node/pretrain_step00053908.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c207fe74dedd0e132c87453b55dbf0777a97249250b9105496c406ad99f43932
+size 7328309140