Qabbel
Phi Tiny MoE Instruct vs Phi-4
Qiegħed żewġ mudelli sa erbgħa ma' xulxin fuq l-oqsma li jiddeċiedu deployment: kuntest, prezz, reġjun, liċenzja, lingwi, hosting, u x'hu verifikat.
| Phi Tiny MoE Instruct | Phi-4 | |
|---|---|---|
| Fornitur | Microsoft | Microsoft |
| Parametri | 3.8B | 14.66B |
| Tieqa tal-kuntest | Mhux verifikat: il-fornitur ma ddokumentax din id-dikjarazzjoni. | 16k |
| Liċenzja | mit | mit |
| Modalitajiet | text | text |
| Data tar-rilaxx | 23 Ġun 2025 | 11 Diċ 2024 |
| Lingwi | Qawwi: en Adekwat: — | Qawwi: en Adekwat: — |
| Piżijiet miftuħa | iva | iva |
| Kompiti | Raġunament, Spiża baxxa | Raġunament, Sommarizzazzjoni, Spiża baxxa, Irkupru (RAG) |
Fejn żewġ mudelli jvarjaw, din it-tabella turi d-differenza u tieqaf. LLM EU ma tippubblikax badge ta' rebbieħ mingħajr punteġġ dokumentat għal dik il-kompitu, u figura nieqsa tibqa' nieqsa minflok ma ssir żero.
Meta tużah
Phi Tiny MoE Instruct
Saħħiet
- open weights are published on Hugging Face and can be downloaded without a vendor agreement
- mit licence, so commercial self-hosting is permitted without a separate agreement
- at 3.8B parameters it fits on a single accelerator
- the model page lists en
- instruction-tuned checkpoint with a chat template in the repository
Limiti
- 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 published configuration records no context window, so the card carries none
- at 3.8B parameters, the weights fit on a single modest accelerator
Kompiti
Phi-4
Saħħiet
- strong MMLU and maths scores at 14B parameters
- MIT licence without user thresholds
- fits a single 40GB accelerator at BF16
Limiti
- 16K-token context window
- declared English-only; EU-language performance is not documented
- training leans on synthetic data, which narrows domain robustness
Kompiti
Kompiti: Raġunament, Spiża baxxa