LLM EU
Meni
Modeli

Medgemma 4B It

Google DeepMind

Samo v katalogu Odprte uteži

Nazadnje pregledano:

Suverenost in gostovanje

Samo v katalogu

Ta model je v katalogu za referenco. Na LLM EU še ni gostovan in noben lastni endpoint ga tukaj ne streže. Zahteve zanj gredo neposredno k ponudniku, pod lastnimi pogoji tega ponudnika.

Povzetek

Medgemma 4B It je srednje velik odprtokodni vizualno-jezikovni model podjetja Google DeepMind: 4.3B parametrov pod licenco health-ai-developer-foundations (po meri, označeno kot "drugo"). Repozitorij je bil ustvarjen maja 2025 in je zrcaljen na Hugging Face. Sprejema besedilo in slike ter odgovarja z besedilom. Kontrolna točka je nastavljena z navodili in vključuje predlogo za klepet. Objavljena konfiguracija ne navaja kontekstnega okna, zato ga ta kartica ne vsebuje. Kartica je samo za katalog: LLM EU ne poganja tega modela. Založnik na strani modela ne navaja jezikovne pokritosti.

Specifikacije

Ponudnik
Google DeepMind (US)
Kontekstno okno
Ni preverjeno: ponudnik te trditve ni dokumentiral.
Parametri
4.3B
Modalnosti
text, vision
Licenca
other · licence
Odprte uteži
yes
Datum izdaje
19. maj 2025
Naloge
Vid, Povzemanje

Jeziki

Močno: —

Zadostno: —

Močno

    Zadostno

      Ni preverjeno: ponudnik te trditve ni dokumentiral. — a language is listed as strong only where a source claims it; the card links the evidence where one exists.

      Kdaj ga uporabiti

      Prednosti

      • open weights are published on Hugging Face and can be downloaded without a vendor agreement
      • at 4.3B parameters it fits on a single accelerator
      • instruction-tuned checkpoint with a chat template in the repository

      Omejitve

      • catalogue only — LLM EU does not run this model, and nothing on this card implies an LLM EU endpoint
      • no evaluation has been run by LLM EU
      • the health-ai-developer-foundations (other) licence is a custom agreement whose commercial terms were not established from its published text, so commercial serving is not permitted until a human clears it; the router must not route to it
      • the published configuration records no context window, so the card carries none
      • at 4.3B parameters, the weights fit on a single modest accelerator
      • vision checkpoints are heavier to serve than their parameter count alone suggests, because the image encoder and pre-processing run alongside the language model
      • a medical model: outputs are not a diagnosis and this catalogue has not evaluated its clinical behaviour