--- base_model: [] library_name: transformers tags: - mergekit - merge --- # Utopia-20B-exp40-3-bf16 This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit). ## Merge Details ### Merge Method This model was merged using the passthrough merge method. ### Models Merged The following models were included in the merge: * E:/Utopia-20B-V1 ### Configuration The following YAML configuration was used to produce this model: ```yaml # Six splits plus "end game # "D" starts at plus .1 VS D/O proj. # 40 plus. # D/G Block Settings. # split 1: .11 to .61 [.04 G] [4 layers] # split 2: .15 to .41 [.08 G] [4 layers] # split 3: .19 to .35 [.23 G] [12 layers] # split 4: .11 to .41 [.244 G] [4 layers] # split 5: .15 to .3 [.248 G] [4 layers] # spilt 6: .19 to .3 [.256 G] [7 layers] # final [D/G]: .33,.44,.55,.66,.77/.88 [5 layers] slices: - sources: - model: E:/Utopia-20B-V1 layer_range: [0, 61] # conc layers # split 1 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.01 - filter: down_proj value: 0.01 - value: 0.11 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.02 - filter: down_proj value: 0.02 - value: 0.12 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.03 - filter: down_proj value: 0.03 - value: 0.13 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.04 - filter: down_proj value: 0.04 - value: 0.61 # split 2, SURGE D THEN D drop .46, continues @ D .15 (from .13) - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.05 - filter: down_proj value: 0.05 - value: 0.15 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.06 - filter: down_proj value: 0.06 - value: 0.16 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.07 - filter: down_proj value: 0.07 - value: 0.17 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.08 - filter: down_proj value: 0.08 - value: 0.41 # split 3, SURGE D to .41, D drop .21 ... follows .17 previous - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.09 - filter: down_proj value: 0.09 - value: 0.19 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.10 - filter: down_proj value: 0.10 - value: 0.20 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.11 - filter: down_proj value: 0.11 - value: .22 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.12 - filter: down_proj value: 0.12 - value: .24 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.13 - filter: down_proj value: 0.13 - value: .26 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.14 - filter: down_proj value: 0.14 - value: .28 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.15 - filter: down_proj value: 0.15 - value: .30 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.16 - filter: down_proj value: 0.16 - value: .31 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.20 - filter: down_proj value: 0.20 - value: .32 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.21 - filter: down_proj value: 0.21 - value: .33 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.22 - filter: down_proj value: 0.22 - value: .34 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.23 - filter: down_proj value: 0.23 - value: .35 # split 4 , NO SURGE D, "D" down drop of .24 ; reverts to .11 (the very first "D" setting ) - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.24 - filter: down_proj value: 0.24 - value: 0.11 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.241 - filter: down_proj value: 0.241 - value: 0.12 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.242 - filter: down_proj value: 0.243 - value: 0.13 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.244 - filter: down_proj value: 0.244 - value: 0.61 # split 5, D Surge to .61, drop to .15 (following .13) - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.245 - filter: down_proj value: 0.245 - value: 0.15 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.246 - filter: down_proj value: 0.246 - value: 0.16 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.247 - filter: down_proj value: 0.247 - value: 0.17 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.248 - filter: down_proj value: 0.248 - value: 0.41 # split 6, D surge to .41 , then follows .17 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.249 - filter: down_proj value: 0.249 - value: 0.19 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.250 - filter: down_proj value: 0.250 - value: 0.20 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.251 - filter: down_proj value: 0.251 - value: .22 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.252 - filter: down_proj value: 0.252 - value: .24 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.253 - filter: down_proj value: 0.254 - value: .26 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.255 - filter: down_proj value: 0.255 - value: .28 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.256 - filter: down_proj value: 0.256 - value: .30 # O PROJ, DPROJ to .3333 / # end game - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.3333333333333 - filter: down_proj value: 0.3333333333333 - value: 0.3333333333333 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.4444444444444 - filter: down_proj value: 0.4444444444444 - value: 0.4444444444444 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.5555555555555 - filter: down_proj value: 0.5555555555555 - value: 0.5555555555555 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.6666666666666 - filter: down_proj value: 0.6666666666666 - value: 0.6666666666666 - sources: - model: E:/Utopia-20B-V1 layer_range: [61,62] parameters: scale: - filter: o_proj value: 0.777777777777 - filter: down_proj value: 0.777777777777 - value: 0.888888888888 merge_method: passthrough dtype: float16 ```