| models: | |
| - model: djuna/L3.1-gramamax | |
| parameters: | |
| weight: [0.5, 0.6, 0.3, 0.9] | |
| density: 0.9 | |
| gamma: 0.1 | |
| - model: djuna/L3.1-Romes-Ninomos | |
| parameters: | |
| weight: [0.5, 0.4, 0.7, 0.1] | |
| density: 0.9 | |
| gamma: 0.1 | |
| base_model: djuna/L3.1-gramamax | |
| parameters: | |
| normalize: false | |
| tokenizer_source: base | |
| merge_method: breadcrumbs_ties | |
| out_dtype: bfloat16 |