Skip to page content
SAVRN
Search Plan a project

Dataset

mesh-1t-prod-mix

by Chutes chutesai/mesh-1t-prod-mix

Canonical archive of the parallax 1T-token production curriculum, phase-split, EXACTLY as prod runs it.

Rows—
Configurations—
Size4.1 TB
Licenseother
AccessAccess requested at publisher
Monthly Downloads68

Dataset Card

Canonical archive of the parallax 1T-token production curriculum, phase-split, EXACTLY as prod runs it. NOT a text dataset: artifacts are flat llama3 uint32 token bins (EOS 128001 between docs), val bins (doc-level 1/100), plus per-pool mixmanifest.json / poolconfig.json provenance and SHA256SUMS. Files >45GB are stored as.part-NNN chunks (cat name.part- > name to reassemble; per-part sha256 in CHUNKS.sha256). (4a) -> 5e6-8e6 or YaRN (4b). Phase-4 replay packed with document-boundary masking ON. Realized per-source shares, exact token counts, floors, sampling seeds, decontamination (13-gram, stride 4) and every assembly decision/deviation: see MANIFEST.json and phase34mixmanifest.json.…

Excerpt from the card by Chutes, licensed other.

Details

Repository
chutesai/mesh-1t-prod-mix
Publisher
Chutes
Task category
Not stated by the source
Tags
Not stated by the source
Size category
Not stated by the source
Languages
Not stated by the source
Revision
5c5954bc1f11d54b23df649d538485bd68c30f6e
Last updated
2026-09-22

Files

169 files, 4.1 TB in total.

Data38 files · 83.0 KB
Documentation1 file · 2.6 KB
Other129 files · 4.1 TB
Repository1 file · 15.0 KB
Every file
FileTypeSizeSHA-256
MANIFEST.jsonData34.1 KB—
members/web_v1/SHA256SUMS.txtData158 B—
members/web_v1/manifest_web_v1.jsonData484 B—
members/web_v1/mix_manifest.jsonData5.1 KB—
phase1/bb35_half/SHA256SUMS.txtData522 B—
phase1/bb35_half/mix_manifest.jsonData1.1 KB—
phase1/bb8_half/SHA256SUMS.txtData609 B—
phase1/bb8_half/mix_manifest.jsonData1.1 KB—
phase2/SHA256SUMS.txtData242 B—
phase2/mix_manifest.jsonData5.8 KB—
phase3/anneal_ccmath4/mix_manifest.jsonData259 B—
phase3/anneal_dolmino/mix_manifest.jsonData264 B—
phase3/anneal_finemath4/mix_manifest.jsonData262 B—
phase3/anneal_infinibyte/mix_manifest.jsonData261 B—
phase3/anneal_omi2/mix_manifest.jsonData259 B—
phase3/anneal_rqa/mix_manifest.jsonData252 B—
phase3/anneal_stem_sft/mix_manifest.jsonData262 B—
phase3/climbmix_top/SHA256SUMS.txtData694 B—
phase3/climbmix_top/mix_manifest.jsonData265 B—
phase3/climbmix_top/pool_config.jsonData1.2 KB—
phase3/code_hq/SHA256SUMS.txtData694 B—
phase3/code_hq/mix_manifest.jsonData1.8 KB—
phase3/code_hq/pool_config.jsonData2.7 KB—
phase3/edupdf_top_v2/SHA256SUMS.txtData609 B—
phase3/edupdf_top_v2/mix_manifest.jsonData268 B—
phase3/edupdf_top_v2/pool_config.jsonData1.1 KB—
phase34_mix_manifest.jsonData17.7 KB—
phase4/longctx_finepdfs_long/mix_manifest.jsonData265 B—
phase4/longctx_longmino/mix_manifest.jsonData264 B—
phase4/longctx_repos/mix_manifest.jsonData260 B—
phase4/longmino_rex_cwe/SHA256SUMS.txtData704 B—
phase4/longmino_rex_cwe/mix_manifest.jsonData438 B—
phase4/nextlong/SHA256SUMS.txtData696 B—
phase4/nextlong/mix_manifest.jsonData437 B—
phase4/pg19/SHA256SUMS.txtData692 B—
phase4/pg19/mix_manifest.jsonData251 B—
phase4/repofim/SHA256SUMS.txtData695 B—
phase4/repofim/mix_manifest.jsonData257 B—
README.mdDocumentation2.6 KB—
members/anneal_b/train.u32.bin.part-000Other34.4 GB—
members/anneal_b/train.u32.bin.part-001Other34.4 GB—
members/anneal_b/train.u32.bin.part-002Other34.4 GB—
members/web_v1/CHUNKS.sha256Other1.3 KB—
members/web_v1/train.u32.bin.part-000Other34.4 GB—
members/web_v1/train.u32.bin.part-001Other34.4 GB—
members/web_v1/train.u32.bin.part-002Other34.4 GB—
members/web_v1/train.u32.bin.part-003Other34.4 GB—
members/web_v1/train.u32.bin.part-004Other34.4 GB—
members/web_v1/train.u32.bin.part-005Other34.4 GB—
members/web_v1/train.u32.bin.part-006Other34.4 GB—
members/web_v1/train.u32.bin.part-007Other34.4 GB—
members/web_v1/train.u32.bin.part-008Other34.4 GB—
members/web_v1/train.u32.bin.part-009Other34.4 GB—
members/web_v1/train.u32.bin.part-010Other34.4 GB—
members/web_v1/train.u32.bin.part-011Other34.4 GB—
members/web_v1/train.u32.bin.part-012Other34.4 GB—
members/web_v1/train.u32.bin.part-013Other34.4 GB—
members/web_v1/train.u32.bin.part-014Other33.8 GB—
members/web_v1/val.u32.binOther5.2 GB—
phase1/bb35_half/CHUNKS.sha256Other4.1 KB—
phase1/bb35_half/train_shard0.u32.bin.part-000Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-001Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-002Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-003Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-004Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-005Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-006Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-007Other34.4 GB—
phase1/bb35_half/train_shard0.u32.bin.part-008Other12.9 GB—
phase1/bb35_half/train_shard1.u32.bin.part-000Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-001Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-002Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-003Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-004Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-005Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-006Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-007Other34.4 GB—
phase1/bb35_half/train_shard1.u32.bin.part-008Other25.2 GB—
phase1/bb35_half/train_shard2.u32.bin.part-000Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-001Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-002Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-003Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-004Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-005Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-006Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-007Other34.4 GB—
phase1/bb35_half/train_shard2.u32.bin.part-008Other25.3 GB—
phase1/bb35_half/train_shard3.u32.bin.part-000Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-001Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-002Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-003Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-004Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-005Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-006Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-007Other34.4 GB—
phase1/bb35_half/train_shard3.u32.bin.part-008Other25.1 GB—
phase1/bb35_half/val.u32.binOther12.0 GB—
phase1/bb8_half/CHUNKS.sha256Other4.0 KB—
phase1/bb8_half/train_shard0.u32.bin.part-000Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-001Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-002Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-003Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-004Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-005Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-006Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-007Other34.4 GB—
phase1/bb8_half/train_shard0.u32.bin.part-008Other13.2 GB—
phase1/bb8_half/train_shard1.u32.bin.part-000Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-001Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-002Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-003Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-004Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-005Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-006Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-007Other34.4 GB—
phase1/bb8_half/train_shard1.u32.bin.part-008Other25.0 GB—
phase1/bb8_half/train_shard2.u32.bin.part-000Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-001Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-002Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-003Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-004Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-005Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-006Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-007Other34.4 GB—
phase1/bb8_half/train_shard2.u32.bin.part-008Other25.2 GB—
phase1/bb8_half/train_shard3.u32.bin.part-000Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-001Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-002Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-003Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-004Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-005Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-006Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-007Other34.4 GB—
phase1/bb8_half/train_shard3.u32.bin.part-008Other25.1 GB—
phase1/bb8_half/val.u32.binOther12.0 GB—
phase2/CHUNKS.sha256Other2.8 KB—
phase2/train.u32.bin.part-000Other34.4 GB—
phase2/train.u32.bin.part-001Other34.4 GB—
phase2/train.u32.bin.part-002Other34.4 GB—
phase2/train.u32.bin.part-003Other34.4 GB—
phase2/train.u32.bin.part-004Other34.4 GB—
phase2/train.u32.bin.part-005Other34.4 GB—
phase2/train.u32.bin.part-006Other34.4 GB—
phase2/train.u32.bin.part-007Other34.4 GB—
phase2/train.u32.bin.part-008Other34.4 GB—
phase2/train.u32.bin.part-009Other34.4 GB—
phase2/train.u32.bin.part-010Other34.4 GB—
phase2/train.u32.bin.part-011Other34.4 GB—
phase2/train.u32.bin.part-012Other34.4 GB—
phase2/train.u32.bin.part-013Other34.4 GB—
phase2/train.u32.bin.part-014Other34.4 GB—
phase2/train.u32.bin.part-015Other34.4 GB—
phase2/train.u32.bin.part-016Other34.4 GB—
phase2/train.u32.bin.part-017Other34.4 GB—
phase2/train.u32.bin.part-018Other34.4 GB—
phase2/train.u32.bin.part-019Other34.4 GB—
phase2/train.u32.bin.part-020Other34.4 GB—
phase2/train.u32.bin.part-021Other34.4 GB—
phase2/train.u32.bin.part-022Other34.4 GB—
phase2/train.u32.bin.part-023Other34.4 GB—
phase2/train.u32.bin.part-024Other34.4 GB—
phase2/train.u32.bin.part-025Other34.4 GB—
phase2/train.u32.bin.part-026Other34.4 GB—
phase2/train.u32.bin.part-027Other34.4 GB—
phase2/train.u32.bin.part-028Other34.4 GB—
phase2/train.u32.bin.part-029Other34.4 GB—
phase2/train.u32.bin.part-030Other2.3 GB—
phase2/val.u32.binOther10.5 GB—
.gitattributesRepository15.0 KB—

License and Download

License
other
Access
Access requested at publisher
Request access from Chutes

Chutes grants access through its official repository on Hugging Face.