Model page

deepseek-ai/deepseek-vl2-tiny warn

downloads 777.9klikes 249license otherarch deepseek_vl_v2params 3370.5Mupdated 2024-12-18

chat template: not found · view on Hugging Face ↗

Scan coverage

Ingot runs three batteries against a model. What each one checks →

BatteryLooks atStatus
Static batteryMetadata & packagingcomplete 2026-08-22
Weights batteryWeights forensics — no GPU, no downloadcomplete 2026-08-22129,280-token embedding scanned · 504 undertrained
Behavioral batteryLive-inference differentialsnot run

Findings

Scanned 2026-08-22 · published from a community scan.

medium Undertrained (glitch) token surface in vocabulary

Embedding-norm scan flagged 504 undertrained tokens (norm < 0.3× the vocabulary median of 9.499), including 24 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "kabungtor", "ultatua", "unisipyo", "jeftigelse", "Kadaghan", "pagklas", "DelaL", "nahimut". In models where this class was tested behaviorally, such tokens silently rewrote user input into confident, schema-valid, wrong output. These are candidates from the weights alone; behavioral confirmation requires the behavioral battery.

How to fixruntime guardweight-level

Keep the affected token strings out of the model's input — the scan-derived runtime guard carries this model's exact blocklist.

  1. Fetch this model's guard artifact (`/api/v1/guard/<owner>/<model>`): the confirmed corrupting tokens and the low-norm candidate list, derived from the published scan.
  2. Screen inbound text with it (the `@ingotai/guard` package is a reference implementation) and route flagged records to a different model or human review — verbatim-copy tasks on flagged strings are the failure mode.
  3. The underlying cause is undertrained embeddings in the weights; a true fix is weight-level (continued pretraining on the affected tokens) — that is not a patch, it's a training job.

Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.

Fingerprint

The durable profile of this model: measured weights-and-metadata facts, rebuilt on every scan and battery run. Updated 2026-08-22.

architecturedeepseek_vl_v2
parameters3370.5M
vocabulary129,280 tokens
licenseother
serializationsafetensors
chat templatenone
glitch-token surface504 undertrained candidates, 24 plain-ASCII
Full measured fingerprint
librarytransformers
pipelineimage-text-to-text
repo files9
revision66c54660eae7
HF snapshot777.9k downloads · 248 likes · updated 2024-12-18 · captured 2026-08-21
embedding tensorlanguage.model.embed_tokens.weight · BF16 · 129,280×1280
embedding normsmedian 9.4993 · mean 9.4107
lineage checkno claimed base model
glitch-token samples"kabungtor", "ultatua", "unisipyo", "jeftigelse", "Kadaghan", "pagklas", "DelaL", "nahimut", "bingkil", "ordenatuak", "asadpan", "asarangang"

Battery runs

The run trace behind the findings above: every deep-battery job for this model, with what each run measured or why it failed. Findings are only as good as the runs that produced them.

batterystatusqueueddurationattempts
weightscomplete2026-08-21 07:4320s1
weights run 2026-08-21 measurements
probes runglitch-norm-scan, lineage-norm-correlation
embedding tensorlanguage.model.embed_tokens.weight · BF16 · 129,280×1280
glitch surface504 undertrained, 24 plain-ASCII
lineage checknot checked (no claimed base model)

Verdict badge

Ship the verdict in your README — it always shows the latest published analysis:

Ingot verdict: warn

[![Ingot scan](https://ingot.tools/api/v1/models/deepseek-ai/deepseek-vl2-tiny/badge.svg)](https://ingot.tools/models/deepseek-ai/deepseek-vl2-tiny)
Gate it in CI