Emhotob — Arabic Translation Scaling Study
From-scratch Arabic LMs (0.5M-50M) finetuned into bidirectional translators across 5 language pairs. Identical recipe at every rung; v1 and v2 bases.
Text Generation • 518k • Updated • 18Note Base · v1 · 0.52M params
oddadmix/Emhotob-1M
Text Generation • 1.07M • Updated • 18Note Base · v1 · 1.08M params
oddadmix/Emhotob-5M
Text Generation • 5.08M • Updated • 18Note Base · v1 · 5.16M params
oddadmix/Emhotob-10M
Text Generation • 10.9M • Updated • 17Note Base · v1 · 11.21M params
oddadmix/Emhotob-25M
Text Generation • 25.3M • Updated • 22Note Base · v1 · 25.3M params
oddadmix/50M-2048-Emhotob
Text Generation • 51.8M • Updated • 933 • 7Note Base · v1 · 51.8M params
oddadmix/Emhotob-500K-v2
Text Generation • 518k • Updated • 26Note Base · v2 · 0.52M params
oddadmix/Emhotob-1M-v2
Text Generation • 1.07M • Updated • 38Note Base · v2 · 1.08M params
oddadmix/Emhotob-5M-v2
Text Generation • 5.08M • Updated • 54Note Base · v2 · 5.16M params
oddadmix/Emhotob-10M-v2
Text Generation • 10.9M • Updated • 17Note Base · v2 · 10.9M params
oddadmix/Emhotob-25M-v2
Text Generation • 25.3M • Updated • 31Note Base · v2 · 25.3M params
oddadmix/Emhotob-500K-MSA-Egyptian-v1
Text Generation • 518k • Updated • 12Note SFT · MSA ↔ Egyptian · 500K v1 · 0.52M params
oddadmix/Emhotob-500K-MSA-Egyptian-v2
Text Generation • 518k • Updated • 13Note SFT · MSA ↔ Egyptian · 500K v2 · 0.52M params
oddadmix/Emhotob-1M-MSA-Egyptian-v1
Text Generation • 1.07M • Updated • 14Note SFT · MSA ↔ Egyptian · 1M v1 · 1.08M params
oddadmix/Emhotob-1M-MSA-Egyptian-v2
Text Generation • 1.07M • Updated • 19Note SFT · MSA ↔ Egyptian · 1M v2 · 1.08M params
oddadmix/Emhotob-5M-MSA-Egyptian-v1
Text Generation • 5.08M • Updated • 11Note SFT · MSA ↔ Egyptian · 5M v1 · 5.16M params
oddadmix/Emhotob-5M-Egyptian-English-v1
Text Generation • 5.08M • Updated • 12Note SFT · English ↔ Egyptian · 5M v1 · 5.16M params
oddadmix/Emhotob-5M-English-MSA-v1
Text Generation • 5.08M • Updated • 7Note SFT · English ↔ MSA · 5M v1 · 5.16M params
oddadmix/Emhotob-5M-Darija-MSA-v1
Text Generation • 5.08M • Updated • 10Note SFT · Darija ↔ MSA · 5M v1 · 5.16M params
oddadmix/Emhotob-5M-Darija-English-v1
Text Generation • 5.08M • Updated • 9Note SFT · Darija ↔ English · 5M v1 · 5.16M params
oddadmix/Emhotob-5M-MSA-Egyptian-v2
Text Generation • 5.08M • Updated • 11Note SFT · MSA ↔ Egyptian · 5M v2 · 5.16M params
oddadmix/Emhotob-5M-Egyptian-English-v2
Text Generation • 5.08M • Updated • 9Note SFT · English ↔ Egyptian · 5M v2 · 5.16M params
oddadmix/Emhotob-5M-English-MSA-v2
Text Generation • 5.08M • Updated • 7Note SFT · English ↔ MSA · 5M v2 · 5.16M params
oddadmix/Emhotob-5M-Darija-MSA-v2
Text Generation • 5.08M • Updated • 14Note SFT · Darija ↔ MSA · 5M v2 · 5.16M params
oddadmix/Emhotob-5M-Darija-English-v2
Text Generation • 5.08M • Updated • 11Note SFT · Darija ↔ English · 5M v2 · 5.16M params
oddadmix/Emhotob-10M-MSA-Egyptian-v1
Text Generation • 10.9M • Updated • 17Note SFT · MSA ↔ Egyptian · 10M v1 · 11.21M params
oddadmix/Emhotob-10M-Egyptian-English-v1
Text Generation • 10.9M • Updated • 13Note SFT · English ↔ Egyptian · 10M v1 · 11.21M params
oddadmix/Emhotob-10M-English-MSA-v1
Text Generation • 10.9M • Updated • 10Note SFT · English ↔ MSA · 10M v1 · 11.21M params
oddadmix/Emhotob-10M-Darija-MSA-v1
Text Generation • 10.9M • Updated • 10Note SFT · Darija ↔ MSA · 10M v1 · 11.21M params
oddadmix/Emhotob-10M-Darija-English-v1
Text Generation • 10.9M • Updated • 10Note SFT · Darija ↔ English · 10M v1 · 11.21M params
oddadmix/Emhotob-10M-MSA-Egyptian-v2
Text Generation • 10.9M • Updated • 18Note SFT · MSA ↔ Egyptian · 10M v2 · 10.9M params
oddadmix/Emhotob-10M-Egyptian-English-v2
Text Generation • 10.9M • Updated • 11Note SFT · English ↔ Egyptian · 10M v2 · 10.9M params
oddadmix/Emhotob-10M-English-MSA-v2
Text Generation • 10.9M • Updated • 5Note SFT · English ↔ MSA · 10M v2 · 10.9M params
oddadmix/Emhotob-10M-Darija-MSA-v2
Text Generation • 10.9M • Updated • 9Note SFT · Darija ↔ MSA · 10M v2 · 10.9M params
oddadmix/Emhotob-10M-Darija-English-v2
Text Generation • 10.9M • Updated • 11Note SFT · Darija ↔ English · 10M v2 · 10.9M params
oddadmix/Emhotob-25M-MSA-Egyptian-v1
Text Generation • 25.3M • Updated • 11Note SFT · MSA ↔ Egyptian · 25M v1 · 25.3M params
oddadmix/Emhotob-25M-Egyptian-English-v1
Text Generation • 25.3M • Updated • 11Note SFT · English ↔ Egyptian · 25M v1 · 25.3M params
oddadmix/Emhotob-25M-English-MSA-v1
Text Generation • 25.3M • Updated • 10Note SFT · English ↔ MSA · 25M v1 · 25.3M params
oddadmix/Emhotob-25M-Darija-MSA-v1
Text Generation • 25.3M • Updated • 13Note SFT · Darija ↔ MSA · 25M v1 · 25.3M params
oddadmix/Emhotob-25M-Darija-English-v1
Text Generation • 25.3M • Updated • 23Note SFT · Darija ↔ English · 25M v1 · 25.3M params
oddadmix/Emhotob-25M-MSA-Egyptian-v2
Text Generation • 25.3M • Updated • 12Note SFT · MSA ↔ Egyptian · 25M v2 · 25.3M params
oddadmix/Emhotob-25M-Egyptian-English-v2
Text Generation • 25.3M • Updated • 16Note SFT · English ↔ Egyptian · 25M v2 · 25.3M params
oddadmix/50M-MSA-Egyptian-v1
Text Generation • 51.8M • Updated • 28Note SFT · MSA ↔ Egyptian · 50M v1 · 51.8M params
oddadmix/50M-Egyptian-English-v1
Text Generation • 51.8M • Updated • 14Note SFT · English ↔ Egyptian · 50M v1 · 51.8M params
oddadmix/50M-English-MSA-v1
Text Generation • 51.8M • Updated • 66Note SFT · English ↔ MSA · 50M v1 · 51.8M params
oddadmix/50M-Darija-MSA-v1
Text Generation • 51.8M • Updated • 11Note SFT · Darija ↔ MSA · 50M v1 · 51.8M params
oddadmix/50M-Darija-English-v1
Text Generation • 51.8M • Updated • 98Note SFT · Darija ↔ English · 50M v1 · 51.8M params
oddadmix/50M-Egyptian-Translation-v1
Text Generation • 51.8M • Updated • 29Note SFT · English → Egyptian (unidirectional) · 50M v1 · 51.8M params