the x1.5 branch's own plan: 945M tokens trained

README.md

@@ -19,7 +19,7 @@ A small language model trained **from scratch**. **Pretraining data and tokenize
 | dataset | 181,924 pages exported 2026-10-07T00:58:27+00:00; after cleaning 142,428 pages: 96,801 crawler pages from 2733 domains + 45,627 record pages |
 | dataset manifest sha256 | `6792bfb740b3c7217663afdba60ae75f5ae6d79aa53ac1475971fabe51a981ac` |
 | unique training tokens | 183,061,543 (+3,726,648 held out); the record pages are about 10,830,251 tokens of it (6%) |
-| tokens trained | 635,961,344 = 3.47 epochs, 6.52 tokens per non-embedding param |
+| tokens trained | 945,029,120 = 5.16 epochs, 9.69 tokens per non-embedding param |
 | tokenizer | BPE, 16384 tokens, trained on this dataset's pages only (crawler pages and record pages) |
 | validation loss | 1.4380 bits/byte on held-out pages (pretraining) |
 | chat tuning (SFT) | `synthetic+openbook`: 1,016,730 question/answer pairs **generated by a general-purpose LLM** from 139,073 of 188,828 page chunks (74%; crawler pages and record pages), plus the pairs cut from the pages. The generator was told to use only facts on each page, and a pair whose key words are not on its page was dropped, but its wording is in this model's chat behaviour. `headings`: 521,532 question/answer pairs cut out of the crawler pages themselves (FAQ questions found on pages, page headings, page titles; answers are verbatim page text). No external model was used. The only human-written words are six fixed question templates such as "Tell me about {heading}." Open book: the generated pairs were also trained once more with the page text they were written from in front of the question (919,976 conversations, 35% with a second page the search ranked high for the same question), so she can answer from the pages retrieved for a question. |

SHA256SUMS

@@ -1,14 +1,14 @@
-12f34705273f5e18a83362b33a0ad06a5449862f27109836f0b2b98664bf2a80  README.md
+bb63a8cab0c38f7c6ad2745d4cbcb8cd94e7bc2671c673656125144b57d243c7  README.md
 a3abb5028b28c5012f2235463bc4dcd65fc114a49da1a569df63b47b41c3aea9  chat.py
 cf2b9f63eb4fe22f63e9eae3f003d5cc76d8553625e234fd2842b8f3df77f7b7  config.json
 dd2a7e32c3c9f8304dcd3c429766071ce6beffda0955beccb8c0a39a05ed82ec  contributors.json
 469a4bf05c641394d550ab653366677ac2102cadc44306f87b5d561daf68a92b  dataset.json
 6792bfb740b3c7217663afdba60ae75f5ae6d79aa53ac1475971fabe51a981ac  dataset_manifest.json
 d954059161546f5ed9301c6ef41fc16423b6cf2f48d9006cdb40961e71bc40e6  model.safetensors
-380a3fab9bc00258435776e73b22f2c7b0a9847bdd84913a2179bfcf24895635  plan.json
+56dc4d56d444f5fcbb2f90daff3f0e0bfd3c7d077b47a56d65157e911b3337b5  plan.json
 b72dd0b69ba2e4171ca100a3d32dab01d26f891a82ce0c8c9c64f945e5ef9d8a  samples.json
 538a27f7191a12cbeed86b768a8ca234b55e9bce697157e1564b097fcdabbcb3  samples.md
 43a9a1532ce197532b7c0824f653a0db43f7c714ce7d3185b4e7e363a82758c7  tokenizer.json
 9a3eac141d0a43f2cb69e64a59757809ff83d49072d70678d61b5caee99fee8b  tokenizer.tiktoken
 99a32dda1542ea57f10a2302a073aa5e1cd657e407a8d199ba39942e1130934f  train_summary.json
-2e8fa113765a0cc4e9bccdc89d770ff02965c2e1e67cbf2b85c39c5bc18405d5  version.json
+c448449309d41dcca959d986c0575e6a41300208242bd6aa7354f2558234561b  version.json

plan.json

@@ -10,29 +10,29 @@
   "unique_train_tokens": 183061543,
   "target_tokens_ratio": 1023939000,
   "target_tokens_epochs": 732246172,
-  "horizon": "epochs",
-  "num_iterations": 2426,
+  "horizon": "branch x1.5 of the stable run",
+  "num_iterations": 3605,
   "total_batch_size": 262144,
   "device_batch_size": 4,
   "world_size": 1,
   "device_batch_size_source": "ladder.json device_batch_size_by_vram",
   "grad_accum_steps": 32,
-  "train_tokens": 635961344,
-  "epochs": 3.47,
-  "tokens_per_param": 6.52,
-  "train_flops": 516146582928752640,
+  "train_tokens": 945029120,
+  "epochs": 5.16,
+  "tokens_per_param": 9.69,
+  "train_flops": 818260411353661440,
   "window_pattern": "L",
   "attention": "sdpa",
   "fp8": false,
   "eval_tokens": 3719168,
-  "eval_every": 121,
-  "save_every": 349,
+  "eval_every": 90,
+  "save_every": 180,
   "keep_checkpoints": true,
   "base_overrides": {},
-  "warmdown_ratio": 0.199861,
+  "warmdown_ratio": 0.2,
   "stream": true,
-  "decay_at": 1941,
-  "plain_at": 2183,
+  "decay_at": 2884,
+  "plain_at": 3245,
   "loader_tails": true,
   "regularization": {
     "wd_schedule": "floor",
@@ -42,10 +42,31 @@
     "dropout_from": 0.5
   },
   "dropout_from_step": null,
-  "wd_horizon": null,
+  "wd_horizon": 3846,
   "save_steps": [],
   "seed": null,
-  "arm": null,
-  "device_type": "cuda",
-  "frozen_at": "2026-10-07T04:28:07"
+  "arm": "R1",
+  "branches": {
+    "1": {
+      "decay_at": 1923,
+      "decay_steps": 481,
+      "num_iterations": 2404
+    },
+    "1.5": {
+      "decay_at": 2884,
+      "decay_steps": 721,
+      "num_iterations": 3605
+    },
+    "2": {
+      "decay_at": 3846,
+      "decay_steps": 962,
+      "num_iterations": 4808
+    }
+  },
+  "tag": "queen-v4-r1-x1.5",
+  "branch": {
+    "of": "queen-v4-r1-stable",
+    "at": 2884,
+    "decay_steps": 721
+  }
 }

version.json

@@ -7,8 +7,8 @@
   "weightsUrl": "https://huggingface.co/Crawlnet/queen-v4",
   "name": "queen-v4",
   "paramsNonEmbedding": 97518000,
-  "tokensTrained": 635961344,
-  "epochs": 3.47,
+  "tokensTrained": 945029120,
+  "epochs": 5.16,
   "valBpb": 1.4380317966698843,
   "sftMethod": "synthetic+openbook",
   "datasetManifestSha256": "6792bfb740b3c7217663afdba60ae75f5ae6d79aa53ac1475971fabe51a981ac",