Model bazowy Poziomka 4B A1.3B. Widział około 80 miliardów tokenów.
Merge modeli z treningów 8,9,10,11
Połączone zostało 47 różnych checkpointów.
Poziomka Baza 8 - https://huggingface.co/cpral/poziomka_8_cpt_hf (prywatne z uwagi na HF storage, mogę upublicznić na prośbę)
Poziomka Baza 9 - https://huggingface.co/cpral/poziomka_9_cpt_hf (prywatne z uwagi na HF storage, mogę upublicznić na prośbę)
Poziomka Baza 10 - https://huggingface.co/cpral/poziomka_10_cpt_hf
Poziomka Baza 11 - https://huggingface.co/cpral/poziomka_11_cpt_hf
Konfiguracja Mergekit
merge_method: linear
models:
- model: /path/poziomka_8_cpt_hf/hf_iter_0000800/model-00001-of-00001.safetensors
parameters:
weight: 0.097129
- model: /path/poziomka_8_cpt_hf/hf_iter_0001600/model-00001-of-00001.safetensors
parameters:
weight: 0.040232
- model: /path/poziomka_8_cpt_hf/hf_iter_0002400/model-00001-of-00001.safetensors
parameters:
weight: 0.030871
- model: /path/poziomka_8_cpt_hf/hf_iter_0003200/model-00001-of-00001.safetensors
parameters:
weight: 0.026026
- model: /path/poziomka_8_cpt_hf/hf_iter_0004000/model-00001-of-00001.safetensors
parameters:
weight: 0.022929
- model: /path/poziomka_8_cpt_hf/hf_iter_0004800/model-00001-of-00001.safetensors
parameters:
weight: 0.020729
- model: /path/poziomka_8_cpt_hf/hf_iter_0005600/model-00001-of-00001.safetensors
parameters:
weight: 0.019063
- model: /path/poziomka_8_cpt_hf/hf_iter_0006400/model-00001-of-00001.safetensors
parameters:
weight: 0.017743
- model: /path/poziomka_8_cpt_hf/hf_iter_0007200/model-00001-of-00001.safetensors
parameters:
weight: 0.016665
- model: /path/poziomka_8_cpt_hf/hf_iter_0008000/model-00001-of-00001.safetensors
parameters:
weight: 0.015762
- model: /path/poziomka_8_cpt_hf/hf_iter_0008800/model-00001-of-00001.safetensors
parameters:
weight: 0.007585
- model: /path/poziomka_8_cpt_hf/hf_iter_0009600/model-00001-of-00001.safetensors
parameters:
weight: 0.007406
- model: /path/poziomka_8_cpt_hf/hf_iter_0010400/model-00001-of-00001.safetensors
parameters:
weight: 0.007240
- model: /path/poziomka_9_cpt_hf/hf_iter_0000800/model-00001-of-00001.safetensors
parameters:
weight: 0.097129
- model: /path/poziomka_9_cpt_hf/hf_iter_0001600/model-00001-of-00001.safetensors
parameters:
weight: 0.040232
- model: /path/poziomka_9_cpt_hf/hf_iter_0002400/model-00001-of-00001.safetensors
parameters:
weight: 0.030871
- model: /path/poziomka_9_cpt_hf/hf_iter_0003200/model-00001-of-00001.safetensors
parameters:
weight: 0.026026
- model: /path/poziomka_9_cpt_hf/hf_iter_0004000/model-00001-of-00001.safetensors
parameters:
weight: 0.022929
- model: /path/poziomka_9_cpt_hf/hf_iter_0004800/model-00001-of-00001.safetensors
parameters:
weight: 0.020729
- model: /path/poziomka_9_cpt_hf/hf_iter_0005600/model-00001-of-00001.safetensors
parameters:
weight: 0.019063
- model: /path/poziomka_9_cpt_hf/hf_iter_0006400/model-00001-of-00001.safetensors
parameters:
weight: 0.017743
- model: /path/poziomka_9_cpt_hf/hf_iter_0007200/model-00001-of-00001.safetensors
parameters:
weight: 0.016665
- model: /path/poziomka_9_cpt_hf/hf_iter_0008000/model-00001-of-00001.safetensors
parameters:
weight: 0.037993
- model: /path/poziomka_10_cpt_hf/iter_0000500/model-00001-of-00001.safetensors
parameters:
weight: 0.017664
- model: /path/poziomka_10_cpt_hf/iter_0001000/model-00001-of-00001.safetensors
parameters:
weight: 0.017215
- model: /path/poziomka_10_cpt_hf/iter_0001500/model-00001-of-00001.safetensors
parameters:
weight: 0.016798
- model: /path/poziomka_10_cpt_hf/iter_0002000/model-00001-of-00001.safetensors
parameters:
weight: 0.016409
- model: /path/poziomka_10_cpt_hf/iter_0002500/model-00001-of-00001.safetensors
parameters:
weight: 0.016047
- model: /path/poziomka_10_cpt_hf/iter_0003000/model-00001-of-00001.safetensors
parameters:
weight: 0.015708
- model: /path/poziomka_10_cpt_hf/iter_0003500/model-00001-of-00001.safetensors
parameters:
weight: 0.015389
- model: /path/poziomka_10_cpt_hf/iter_0004000/model-00001-of-00001.safetensors
parameters:
weight: 0.015089
- model: /path/poziomka_10_cpt_hf/iter_0004500/model-00001-of-00001.safetensors
parameters:
weight: 0.014806
- model: /path/poziomka_10_cpt_hf/iter_0005000/model-00001-of-00001.safetensors
parameters:
weight: 0.014538
- model: /path/poziomka_10_cpt_hf/iter_0005500/model-00001-of-00001.safetensors
parameters:
weight: 0.014284
- model: /path/poziomka_10_cpt_hf/iter_0006000/model-00001-of-00001.safetensors
parameters:
weight: 0.014043
- model: /path/poziomka_11_cpt_hf/iter_0000500/model-00001-of-00001.safetensors
parameters:
weight: 0.013814
- model: /path/poziomka_11_cpt_hf/iter_0001000/model-00001-of-00001.safetensors
parameters:
weight: 0.013596
- model: /path/poziomka_11_cpt_hf/iter_0001500/model-00001-of-00001.safetensors
parameters:
weight: 0.013388
- model: /path/poziomka_11_cpt_hf/iter_0002000/model-00001-of-00001.safetensors
parameters:
weight: 0.013189
- model: /path/poziomka_11_cpt_hf/iter_0002500/model-00001-of-00001.safetensors
parameters:
weight: 0.012998
- model: /path/poziomka_11_cpt_hf/iter_0003000/model-00001-of-00001.safetensors
parameters:
weight: 0.012816
- model: /path/poziomka_11_cpt_hf/iter_0003500/model-00001-of-00001.safetensors
parameters:
weight: 0.012641
- model: /path/poziomka_11_cpt_hf/iter_0004000/model-00001-of-00001.safetensors
parameters:
weight: 0.012473
- model: /path/poziomka_11_cpt_hf/iter_0004500/model-00001-of-00001.safetensors
parameters:
weight: 0.012312
- model: /path/poziomka_11_cpt_hf/iter_0005000/model-00001-of-00001.safetensors
parameters:
weight: 0.012157
- model: /path/poziomka_11_cpt_hf/iter_0005500/model-00001-of-00001.safetensors
parameters:
weight: 0.012007
- model: /path/poziomka_11_cpt_hf/iter_0006000/model-00001-of-00001.safetensors
parameters:
weight: 0.011863
dtype: bfloat16