Independent publisher
Akshay Kumar Gautam
glouriousgautam
AI
Compare the published evidence
This profile connects the 3 model and dataset records shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.
Datasets
Nine pools across four isolated CPU notebooks. This repository owns: {"educationalweb": 4000000000, "generalweb": 8000000000}. Read latest.json, then its pinned checkpoint manifest. Counts include one EOS per document. Uint16 little-endian token banks retain document offsets and original English text in sibling Parquet indexes. Training performs its own context packing and token-aware source sampling. Index offsets are part-local. Strict natural-language checks, permissive semantic junk screening, local normalized exact deduplication. Cross-pool/prior-training deduplication and semantic/language sample review are explicitly pending. Source quality and language classifiers are fallible. No…
Nine pools across four isolated CPU notebooks. This repository owns: {"code": 6000000000, "json": 500000000, "validatedtools": 500000000}. Read latest.json, then its pinned checkpoint manifest. Counts include one EOS per document. Uint16 little-endian token banks retain document offsets and original English text in sibling Parquet indexes. Training performs its own context packing and token-aware source sampling. Index offsets are part-local. Strict natural-language checks, permissive semantic junk screening, local normalized exact deduplication. Cross-pool/prior-training deduplication and semantic/language sample review are explicitly pending. Source quality and language classifiers are…