Upload folder using huggingface_hub

.ipynb_checkpoints/config-checkpoint.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "hidden_act": "silu",
+  "hidden_size": 8192,
+  "initializer_range": 0.02,
+  "intermediate_size": 28672,
+  "max_position_embeddings": 32768,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 64,
+  "num_hidden_layers": 80,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 3000000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.44.2",
+  "use_cache": true,
+  "use_flash_attention_2": true,
+  "vocab_size": 128256
+}

.ipynb_checkpoints/model.safetensors.index-checkpoint.json

@@ -0,0 +1 @@
+{"metadata": {"mergekit_version": "0.0.4.4", "total_size": 282214825984}, "weight_map": {"lm_head.weight": "model-00001-of-00062.safetensors", "model.embed_tokens.weight": "model-00002-of-00062.safetensors", "model.layers.0.input_layernorm.weight": "model-00002-of-00062.safetensors", "model.layers.0.mlp.down_proj.weight": "model-00003-of-00062.safetensors", "model.layers.0.mlp.gate_proj.weight": "model-00003-of-00062.safetensors", "model.layers.0.mlp.up_proj.weight": "model-00003-of-00062.safetensors", "model.layers.0.post_attention_layernorm.weight": "model-00003-of-00062.safetensors", "model.layers.0.self_attn.k_proj.weight": "model-00003-of-00062.safetensors", "model.layers.0.self_attn.o_proj.weight": "model-00003-of-00062.safetensors", "model.layers.0.self_attn.q_proj.weight": "model-00003-of-00062.safetensors", "model.layers.0.self_attn.v_proj.weight": "model-00003-of-00062.safetensors", "model.layers.1.input_layernorm.weight": "model-00003-of-00062.safetensors", "model.layers.1.mlp.down_proj.weight": "model-00003-of-00062.safetensors", "model.layers.1.mlp.gate_proj.weight": "model-00004-of-00062.safetensors", "model.layers.1.mlp.up_proj.weight": "model-00004-of-00062.safetensors", "model.layers.1.post_attention_layernorm.weight": "model-00004-of-00062.safetensors", "model.layers.1.self_attn.k_proj.weight": "model-00004-of-00062.safetensors", "model.layers.1.self_attn.o_proj.weight": "model-00004-of-00062.safetensors", "model.layers.1.self_attn.q_proj.weight": "model-00004-of-00062.safetensors", "model.layers.1.self_attn.v_proj.weight": "model-00004-of-00062.safetensors", "model.layers.10.input_layernorm.weight": "model-00004-of-00062.safetensors", "model.layers.10.mlp.down_proj.weight": "model-00004-of-00062.safetensors", "model.layers.10.mlp.gate_proj.weight": "model-00004-of-00062.safetensors", "model.layers.10.mlp.up_proj.weight": "model-00005-of-00062.safetensors", "model.layers.10.post_attention_layernorm.weight": "model-00005-of-00062.safetensors", "model.layers.10.self_attn.k_proj.weight": "model-00005-of-00062.safetensors", "model.layers.10.self_attn.o_proj.weight": "model-00005-of-00062.safetensors", "model.layers.10.self_attn.q_proj.weight": "model-00005-of-00062.safetensors", "model.layers.10.self_attn.v_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.input_layernorm.weight": "model-00005-of-00062.safetensors", "model.layers.11.mlp.down_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.mlp.gate_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.mlp.up_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.post_attention_layernorm.weight": "model-00005-of-00062.safetensors", "model.layers.11.self_attn.k_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.self_attn.o_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.self_attn.q_proj.weight": "model-00005-of-00062.safetensors", "model.layers.11.self_attn.v_proj.weight": "model-00005-of-00062.safetensors", "model.layers.12.input_layernorm.weight": "model-00005-of-00062.safetensors", "model.layers.12.mlp.down_proj.weight": "model-00006-of-00062.safetensors", "model.layers.12.mlp.gate_proj.weight": "model-00006-of-00062.safetensors", "model.layers.12.mlp.up_proj.weight": "model-00006-of-00062.safetensors", "model.layers.12.post_attention_layernorm.weight": "model-00006-of-00062.safetensors", "model.layers.12.self_attn.k_proj.weight": "model-00006-of-00062.safetensors", "model.layers.12.self_attn.o_proj.weight": "model-00006-of-00062.safetensors", "model.layers.12.self_attn.q_proj.weight": "model-00006-of-00062.safetensors", "model.layers.12.self_attn.v_proj.weight": "model-00006-of-00062.safetensors", "model.layers.13.input_layernorm.weight": "model-00006-of-00062.safetensors", "model.layers.13.mlp.down_proj.weight": "model-00006-of-00062.safetensors", "model.layers.13.mlp.gate_proj.weight": "model-00007-of-00062.safetensors", "model.layers.13.mlp.up_proj.weight": "model-00007-of-00062.safetensors", "model.layers.13.post_attention_layernorm.weight": "model-00007-of-00062.safetensors", "model.layers.13.self_attn.k_proj.weight": "model-00007-of-00062.safetensors", "model.layers.13.self_attn.o_proj.weight": "model-00007-of-00062.safetensors", "model.layers.13.self_attn.q_proj.weight": "model-00007-of-00062.safetensors", "model.layers.13.self_attn.v_proj.weight": "model-00007-of-00062.safetensors", "model.layers.14.input_layernorm.weight": "model-00007-of-00062.safetensors", "model.layers.14.mlp.down_proj.weight": "model-00007-of-00062.safetensors", "model.layers.14.mlp.gate_proj.weight": "model-00007-of-00062.safetensors", "model.layers.14.mlp.up_proj.weight": "model-00008-of-00062.safetensors", "model.layers.14.post_attention_layernorm.weight": "model-00008-of-00062.safetensors", "model.layers.14.self_attn.k_proj.weight": "model-00008-of-00062.safetensors", "model.layers.14.self_attn.o_proj.weight": "model-00008-of-00062.safetensors", "model.layers.14.self_attn.q_proj.weight": "model-00008-of-00062.safetensors", "model.layers.14.self_attn.v_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.input_layernorm.weight": "model-00008-of-00062.safetensors", "model.layers.15.mlp.down_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.mlp.gate_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.mlp.up_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.post_attention_layernorm.weight": "model-00008-of-00062.safetensors", "model.layers.15.self_attn.k_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.self_attn.o_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.self_attn.q_proj.weight": "model-00008-of-00062.safetensors", "model.layers.15.self_attn.v_proj.weight": "model-00008-of-00062.safetensors", "model.layers.16.input_layernorm.weight": "model-00008-of-00062.safetensors", "model.layers.16.mlp.down_proj.weight": "model-00009-of-00062.safetensors", "model.layers.16.mlp.gate_proj.weight": "model-00009-of-00062.safetensors", "model.layers.16.mlp.up_proj.weight": "model-00009-of-00062.safetensors", "model.layers.16.post_attention_layernorm.weight": "model-00009-of-00062.safetensors", "model.layers.16.self_attn.k_proj.weight": "model-00009-of-00062.safetensors", "model.layers.16.self_attn.o_proj.weight": "model-00009-of-00062.safetensors", "model.layers.16.self_attn.q_proj.weight": "model-00009-of-00062.safetensors", "model.layers.16.self_attn.v_proj.weight": "model-00009-of-00062.safetensors", "model.layers.17.input_layernorm.weight": "model-00009-of-00062.safetensors", "model.layers.17.mlp.down_proj.weight": "model-00009-of-00062.safetensors", "model.layers.17.mlp.gate_proj.weight": "model-00010-of-00062.safetensors", "model.layers.17.mlp.up_proj.weight": "model-00010-of-00062.safetensors", "model.layers.17.post_attention_layernorm.weight": "model-00010-of-00062.safetensors", "model.layers.17.self_attn.k_proj.weight": "model-00010-of-00062.safetensors", "model.layers.17.self_attn.o_proj.weight": "model-00010-of-00062.safetensors", "model.layers.17.self_attn.q_proj.weight": "model-00010-of-00062.safetensors", "model.layers.17.self_attn.v_proj.weight": "model-00010-of-00062.safetensors", "model.layers.18.input_layernorm.weight": "model-00010-of-00062.safetensors", "model.layers.18.mlp.down_proj.weight": "model-00010-of-00062.safetensors", "model.layers.18.mlp.gate_proj.weight": "model-00010-of-00062.safetensors", "model.layers.18.mlp.up_proj.weight": "model-00011-of-00062.safetensors", "model.layers.18.post_attention_layernorm.weight": "model-00011-of-00062.safetensors", "model.layers.18.self_attn.k_proj.weight": "model-00011-of-00062.safetensors", "model.layers.18.self_attn.o_proj.weight": "model-00011-of-00062.safetensors", "model.layers.18.self_attn.q_proj.weight": "model-00011-of-00062.safetensors", "model.layers.18.self_attn.v_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.input_layernorm.weight": "model-00011-of-00062.safetensors", "model.layers.19.mlp.down_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.mlp.gate_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.mlp.up_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.post_attention_layernorm.weight": "model-00011-of-00062.safetensors", "model.layers.19.self_attn.k_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.self_attn.o_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.self_attn.q_proj.weight": "model-00011-of-00062.safetensors", "model.layers.19.self_attn.v_proj.weight": "model-00011-of-00062.safetensors", "model.layers.2.input_layernorm.weight": "model-00011-of-00062.safetensors", "model.layers.2.mlp.down_proj.weight": "model-00012-of-00062.safetensors", "model.layers.2.mlp.gate_proj.weight": "model-00012-of-00062.safetensors", "model.layers.2.mlp.up_proj.weight": "model-00012-of-00062.safetensors", "model.layers.2.post_attention_layernorm.weight": "model-00012-of-00062.safetensors", "model.layers.2.self_attn.k_proj.weight": "model-00012-of-00062.safetensors", "model.layers.2.self_attn.o_proj.weight": "model-00012-of-00062.safetensors", "model.layers.2.self_attn.q_proj.weight": "model-00012-of-00062.safetensors", "model.layers.2.self_attn.v_proj.weight": "model-00012-of-00062.safetensors", "model.layers.20.input_layernorm.weight": "model-00012-of-00062.safetensors", "model.layers.20.mlp.down_proj.weight": "model-00012-of-00062.safetensors", "model.layers.20.mlp.gate_proj.weight": "model-00013-of-00062.safetensors", "model.layers.20.mlp.up_proj.weight": "model-00013-of-00062.safetensors", "model.layers.20.post_attention_layernorm.weight": "model-00013-of-00062.safetensors", "model.layers.20.self_attn.k_proj.weight": "model-00013-of-00062.safetensors", "model.layers.20.self_attn.o_proj.weight": "model-00013-of-00062.safetensors", "model.layers.20.self_attn.q_proj.weight": "model-00013-of-00062.safetensors", "model.layers.20.self_attn.v_proj.weight": "model-00013-of-00062.safetensors", "model.layers.21.input_layernorm.weight": "model-00013-of-00062.safetensors", "model.layers.21.mlp.down_proj.weight": "model-00013-of-00062.safetensors", "model.layers.21.mlp.gate_proj.weight": "model-00013-of-00062.safetensors", "model.layers.21.mlp.up_proj.weight": "model-00014-of-00062.safetensors", "model.layers.21.post_attention_layernorm.weight": "model-00014-of-00062.safetensors", "model.layers.21.self_attn.k_proj.weight": "model-00014-of-00062.safetensors", "model.layers.21.self_attn.o_proj.weight": "model-00014-of-00062.safetensors", "model.layers.21.self_attn.q_proj.weight": "model-00014-of-00062.safetensors", "model.layers.21.self_attn.v_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.input_layernorm.weight": "model-00014-of-00062.safetensors", "model.layers.22.mlp.down_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.mlp.gate_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.mlp.up_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.post_attention_layernorm.weight": "model-00014-of-00062.safetensors", "model.layers.22.self_attn.k_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.self_attn.o_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.self_attn.q_proj.weight": "model-00014-of-00062.safetensors", "model.layers.22.self_attn.v_proj.weight": "model-00014-of-00062.safetensors", "model.layers.23.input_layernorm.weight": "model-00014-of-00062.safetensors", "model.layers.23.mlp.down_proj.weight": "model-00015-of-00062.safetensors", "model.layers.23.mlp.gate_proj.weight": "model-00015-of-00062.safetensors", "model.layers.23.mlp.up_proj.weight": "model-00015-of-00062.safetensors", "model.layers.23.post_attention_layernorm.weight": "model-00015-of-00062.safetensors", "model.layers.23.self_attn.k_proj.weight": "model-00015-of-00062.safetensors", "model.layers.23.self_attn.o_proj.weight": "model-00015-of-00062.safetensors", "model.layers.23.self_attn.q_proj.weight": "model-00015-of-00062.safetensors", "model.layers.23.self_attn.v_proj.weight": "model-00015-of-00062.safetensors", "model.layers.24.input_layernorm.weight": "model-00015-of-00062.safetensors", "model.layers.24.mlp.down_proj.weight": "model-00015-of-00062.safetensors", "model.layers.24.mlp.gate_proj.weight": "model-00016-of-00062.safetensors", "model.layers.24.mlp.up_proj.weight": "model-00016-of-00062.safetensors", "model.layers.24.post_attention_layernorm.weight": "model-00016-of-00062.safetensors", "model.layers.24.self_attn.k_proj.weight": "model-00016-of-00062.safetensors", "model.layers.24.self_attn.o_proj.weight": "model-00016-of-00062.safetensors", "model.layers.24.self_attn.q_proj.weight": "model-00016-of-00062.safetensors", "model.layers.24.self_attn.v_proj.weight": "model-00016-of-00062.safetensors", "model.layers.25.input_layernorm.weight": "model-00016-of-00062.safetensors", "model.layers.25.mlp.down_proj.weight": "model-00016-of-00062.safetensors", "model.layers.25.mlp.gate_proj.weight": "model-00016-of-00062.safetensors", "model.layers.25.mlp.up_proj.weight": "model-00017-of-00062.safetensors", "model.layers.25.post_attention_layernorm.weight": "model-00017-of-00062.safetensors", "model.layers.25.self_attn.k_proj.weight": "model-00017-of-00062.safetensors", "model.layers.25.self_attn.o_proj.weight": "model-00017-of-00062.safetensors", "model.layers.25.self_attn.q_proj.weight": "model-00017-of-00062.safetensors", "model.layers.25.self_attn.v_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.input_layernorm.weight": "model-00017-of-00062.safetensors", "model.layers.26.mlp.down_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.mlp.gate_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.mlp.up_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.post_attention_layernorm.weight": "model-00017-of-00062.safetensors", "model.layers.26.self_attn.k_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.self_attn.o_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.self_attn.q_proj.weight": "model-00017-of-00062.safetensors", "model.layers.26.self_attn.v_proj.weight": "model-00017-of-00062.safetensors", "model.layers.27.input_layernorm.weight": "model-00017-of-00062.safetensors", "model.layers.27.mlp.down_proj.weight": "model-00018-of-00062.safetensors", "model.layers.27.mlp.gate_proj.weight": "model-00018-of-00062.safetensors", "model.layers.27.mlp.up_proj.weight": "model-00018-of-00062.safetensors", "model.layers.27.post_attention_layernorm.weight": "model-00018-of-00062.safetensors", "model.layers.27.self_attn.k_proj.weight": "model-00018-of-00062.safetensors", "model.layers.27.self_attn.o_proj.weight": "model-00018-of-00062.safetensors", "model.layers.27.self_attn.q_proj.weight": "model-00018-of-00062.safetensors", "model.layers.27.self_attn.v_proj.weight": "model-00018-of-00062.safetensors", "model.layers.28.input_layernorm.weight": "model-00018-of-00062.safetensors", "model.layers.28.mlp.down_proj.weight": "model-00018-of-00062.safetensors", "model.layers.28.mlp.gate_proj.weight": "model-00019-of-00062.safetensors", "model.layers.28.mlp.up_proj.weight": "model-00019-of-00062.safetensors", "model.layers.28.post_attention_layernorm.weight": "model-00019-of-00062.safetensors", "model.layers.28.self_attn.k_proj.weight": "model-00019-of-00062.safetensors", "model.layers.28.self_attn.o_proj.weight": "model-00019-of-00062.safetensors", "model.layers.28.self_attn.q_proj.weight": "model-00019-of-00062.safetensors", "model.layers.28.self_attn.v_proj.weight": "model-00019-of-00062.safetensors", "model.layers.29.input_layernorm.weight": "model-00019-of-00062.safetensors", "model.layers.29.mlp.down_proj.weight": "model-00019-of-00062.safetensors", "model.layers.29.mlp.gate_proj.weight": "model-00019-of-00062.safetensors", "model.layers.29.mlp.up_proj.weight": "model-00020-of-00062.safetensors", "model.layers.29.post_attention_layernorm.weight": "model-00020-of-00062.safetensors", "model.layers.29.self_attn.k_proj.weight": "model-00020-of-00062.safetensors", "model.layers.29.self_attn.o_proj.weight": "model-00020-of-00062.safetensors", "model.layers.29.self_attn.q_proj.weight": "model-00020-of-00062.safetensors", "model.layers.29.self_attn.v_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.input_layernorm.weight": "model-00020-of-00062.safetensors", "model.layers.3.mlp.down_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.mlp.gate_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.mlp.up_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.post_attention_layernorm.weight": "model-00020-of-00062.safetensors", "model.layers.3.self_attn.k_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.self_attn.o_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.self_attn.q_proj.weight": "model-00020-of-00062.safetensors", "model.layers.3.self_attn.v_proj.weight": "model-00020-of-00062.safetensors", "model.layers.30.input_layernorm.weight": "model-00020-of-00062.safetensors", "model.layers.30.mlp.down_proj.weight": "model-00021-of-00062.safetensors", "model.layers.30.mlp.gate_proj.weight": "model-00021-of-00062.safetensors", "model.layers.30.mlp.up_proj.weight": "model-00021-of-00062.safetensors", "model.layers.30.post_attention_layernorm.weight": "model-00021-of-00062.safetensors", "model.layers.30.self_attn.k_proj.weight": "model-00021-of-00062.safetensors", "model.layers.30.self_attn.o_proj.weight": "model-00021-of-00062.safetensors", "model.layers.30.self_attn.q_proj.weight": "model-00021-of-00062.safetensors", "model.layers.30.self_attn.v_proj.weight": "model-00021-of-00062.safetensors", "model.layers.31.input_layernorm.weight": "model-00021-of-00062.safetensors", "model.layers.31.mlp.down_proj.weight": "model-00021-of-00062.safetensors", "model.layers.31.mlp.gate_proj.weight": "model-00022-of-00062.safetensors", "model.layers.31.mlp.up_proj.weight": "model-00022-of-00062.safetensors", "model.layers.31.post_attention_layernorm.weight": "model-00022-of-00062.safetensors", "model.layers.31.self_attn.k_proj.weight": "model-00022-of-00062.safetensors", "model.layers.31.self_attn.o_proj.weight": "model-00022-of-00062.safetensors", "model.layers.31.self_attn.q_proj.weight": "model-00022-of-00062.safetensors", "model.layers.31.self_attn.v_proj.weight": "model-00022-of-00062.safetensors", "model.layers.32.input_layernorm.weight": "model-00022-of-00062.safetensors", "model.layers.32.mlp.down_proj.weight": "model-00022-of-00062.safetensors", "model.layers.32.mlp.gate_proj.weight": "model-00022-of-00062.safetensors", "model.layers.32.mlp.up_proj.weight": "model-00023-of-00062.safetensors", "model.layers.32.post_attention_layernorm.weight": "model-00023-of-00062.safetensors", "model.layers.32.self_attn.k_proj.weight": "model-00023-of-00062.safetensors", "model.layers.32.self_attn.o_proj.weight": "model-00023-of-00062.safetensors", "model.layers.32.self_attn.q_proj.weight": "model-00023-of-00062.safetensors", "model.layers.32.self_attn.v_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.input_layernorm.weight": "model-00023-of-00062.safetensors", "model.layers.33.mlp.down_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.mlp.gate_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.mlp.up_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.post_attention_layernorm.weight": "model-00023-of-00062.safetensors", "model.layers.33.self_attn.k_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.self_attn.o_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.self_attn.q_proj.weight": "model-00023-of-00062.safetensors", "model.layers.33.self_attn.v_proj.weight": "model-00023-of-00062.safetensors", "model.layers.34.input_layernorm.weight": "model-00023-of-00062.safetensors", "model.layers.34.mlp.down_proj.weight": "model-00024-of-00062.safetensors", "model.layers.34.mlp.gate_proj.weight": "model-00024-of-00062.safetensors", "model.layers.34.mlp.up_proj.weight": "model-00024-of-00062.safetensors", "model.layers.34.post_attention_layernorm.weight": "model-00024-of-00062.safetensors", "model.layers.34.self_attn.k_proj.weight": "model-00024-of-00062.safetensors", "model.layers.34.self_attn.o_proj.weight": "model-00024-of-00062.safetensors", "model.layers.34.self_attn.q_proj.weight": "model-00024-of-00062.safetensors", "model.layers.34.self_attn.v_proj.weight": "model-00024-of-00062.safetensors", "model.layers.35.input_layernorm.weight": "model-00024-of-00062.safetensors", "model.layers.35.mlp.down_proj.weight": "model-00024-of-00062.safetensors", "model.layers.35.mlp.gate_proj.weight": "model-00025-of-00062.safetensors", "model.layers.35.mlp.up_proj.weight": "model-00025-of-00062.safetensors", "model.layers.35.post_attention_layernorm.weight": "model-00025-of-00062.safetensors", "model.layers.35.self_attn.k_proj.weight": "model-00025-of-00062.safetensors", "model.layers.35.self_attn.o_proj.weight": "model-00025-of-00062.safetensors", "model.layers.35.self_attn.q_proj.weight": "model-00025-of-00062.safetensors", "model.layers.35.self_attn.v_proj.weight": "model-00025-of-00062.safetensors", "model.layers.36.input_layernorm.weight": "model-00025-of-00062.safetensors", "model.layers.36.mlp.down_proj.weight": "model-00025-of-00062.safetensors", "model.layers.36.mlp.gate_proj.weight": "model-00025-of-00062.safetensors", "model.layers.36.mlp.up_proj.weight": "model-00026-of-00062.safetensors", "model.layers.36.post_attention_layernorm.weight": "model-00026-of-00062.safetensors", "model.layers.36.self_attn.k_proj.weight": "model-00026-of-00062.safetensors", "model.layers.36.self_attn.o_proj.weight": "model-00026-of-00062.safetensors", "model.layers.36.self_attn.q_proj.weight": "model-00026-of-00062.safetensors", "model.layers.36.self_attn.v_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.input_layernorm.weight": "model-00026-of-00062.safetensors", "model.layers.37.mlp.down_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.mlp.gate_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.mlp.up_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.post_attention_layernorm.weight": "model-00026-of-00062.safetensors", "model.layers.37.self_attn.k_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.self_attn.o_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.self_attn.q_proj.weight": "model-00026-of-00062.safetensors", "model.layers.37.self_attn.v_proj.weight": "model-00026-of-00062.safetensors", "model.layers.38.input_layernorm.weight": "model-00026-of-00062.safetensors", "model.layers.38.mlp.down_proj.weight": "model-00027-of-00062.safetensors", "model.layers.38.mlp.gate_proj.weight": "model-00027-of-00062.safetensors", "model.layers.38.mlp.up_proj.weight": "model-00027-of-00062.safetensors", "model.layers.38.post_attention_layernorm.weight": "model-00027-of-00062.safetensors", "model.layers.38.self_attn.k_proj.weight": "model-00027-of-00062.safetensors", "model.layers.38.self_attn.o_proj.weight": "model-00027-of-00062.safetensors", "model.layers.38.self_attn.q_proj.weight": "model-00027-of-00062.safetensors", "model.layers.38.self_attn.v_proj.weight": "model-00027-of-00062.safetensors", "model.layers.39.input_layernorm.weight": "model-00027-of-00062.safetensors", "model.layers.39.mlp.down_proj.weight": "model-00027-of-00062.safetensors", "model.layers.39.mlp.gate_proj.weight": "model-00028-of-00062.safetensors", "model.layers.39.mlp.up_proj.weight": "model-00028-of-00062.safetensors", "model.layers.39.post_attention_layernorm.weight": "model-00028-of-00062.safetensors", "model.layers.39.self_attn.k_proj.weight": "model-00028-of-00062.safetensors", "model.layers.39.self_attn.o_proj.weight": "model-00028-of-00062.safetensors", "model.layers.39.self_attn.q_proj.weight": "model-00028-of-00062.safetensors", "model.layers.39.self_attn.v_proj.weight": "model-00028-of-00062.safetensors", "model.layers.4.input_layernorm.weight": "model-00028-of-00062.safetensors", "model.layers.4.mlp.down_proj.weight": "model-00028-of-00062.safetensors", "model.layers.4.mlp.gate_proj.weight": "model-00028-of-00062.safetensors", "model.layers.4.mlp.up_proj.weight": "model-00029-of-00062.safetensors", "model.layers.4.post_attention_layernorm.weight": "model-00029-of-00062.safetensors", "model.layers.4.self_attn.k_proj.weight": "model-00029-of-00062.safetensors", "model.layers.4.self_attn.o_proj.weight": "model-00029-of-00062.safetensors", "model.layers.4.self_attn.q_proj.weight": "model-00029-of-00062.safetensors", "model.layers.4.self_attn.v_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.input_layernorm.weight": "model-00029-of-00062.safetensors", "model.layers.40.mlp.down_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.mlp.gate_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.mlp.up_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.post_attention_layernorm.weight": "model-00029-of-00062.safetensors", "model.layers.40.self_attn.k_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.self_attn.o_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.self_attn.q_proj.weight": "model-00029-of-00062.safetensors", "model.layers.40.self_attn.v_proj.weight": "model-00029-of-00062.safetensors", "model.layers.41.input_layernorm.weight": "model-00029-of-00062.safetensors", "model.layers.41.mlp.down_proj.weight": "model-00030-of-00062.safetensors", "model.layers.41.mlp.gate_proj.weight": "model-00030-of-00062.safetensors", "model.layers.41.mlp.up_proj.weight": "model-00030-of-00062.safetensors", "model.layers.41.post_attention_layernorm.weight": "model-00030-of-00062.safetensors", "model.layers.41.self_attn.k_proj.weight": "model-00030-of-00062.safetensors", "model.layers.41.self_attn.o_proj.weight": "model-00030-of-00062.safetensors", "model.layers.41.self_attn.q_proj.weight": "model-00030-of-00062.safetensors", "model.layers.41.self_attn.v_proj.weight": "model-00030-of-00062.safetensors", "model.layers.42.input_layernorm.weight": "model-00030-of-00062.safetensors", "model.layers.42.mlp.down_proj.weight": "model-00030-of-00062.safetensors", "model.layers.42.mlp.gate_proj.weight": "model-00031-of-00062.safetensors", "model.layers.42.mlp.up_proj.weight": "model-00031-of-00062.safetensors", "model.layers.42.post_attention_layernorm.weight": "model-00031-of-00062.safetensors", "model.layers.42.self_attn.k_proj.weight": "model-00031-of-00062.safetensors", "model.layers.42.self_attn.o_proj.weight": "model-00031-of-00062.safetensors", "model.layers.42.self_attn.q_proj.weight": "model-00031-of-00062.safetensors", "model.layers.42.self_attn.v_proj.weight": "model-00031-of-00062.safetensors", "model.layers.43.input_layernorm.weight": "model-00031-of-00062.safetensors", "model.layers.43.mlp.down_proj.weight": "model-00031-of-00062.safetensors", "model.layers.43.mlp.gate_proj.weight": "model-00031-of-00062.safetensors", "model.layers.43.mlp.up_proj.weight": "model-00032-of-00062.safetensors", "model.layers.43.post_attention_layernorm.weight": "model-00032-of-00062.safetensors", "model.layers.43.self_attn.k_proj.weight": "model-00032-of-00062.safetensors", "model.layers.43.self_attn.o_proj.weight": "model-00032-of-00062.safetensors", "model.layers.43.self_attn.q_proj.weight": "model-00032-of-00062.safetensors", "model.layers.43.self_attn.v_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.input_layernorm.weight": "model-00032-of-00062.safetensors", "model.layers.44.mlp.down_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.mlp.gate_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.mlp.up_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.post_attention_layernorm.weight": "model-00032-of-00062.safetensors", "model.layers.44.self_attn.k_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.self_attn.o_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.self_attn.q_proj.weight": "model-00032-of-00062.safetensors", "model.layers.44.self_attn.v_proj.weight": "model-00032-of-00062.safetensors", "model.layers.45.input_layernorm.weight": "model-00032-of-00062.safetensors", "model.layers.45.mlp.down_proj.weight": "model-00033-of-00062.safetensors", "model.layers.45.mlp.gate_proj.weight": "model-00033-of-00062.safetensors", "model.layers.45.mlp.up_proj.weight": "model-00033-of-00062.safetensors", "model.layers.45.post_attention_layernorm.weight": "model-00033-of-00062.safetensors", "model.layers.45.self_attn.k_proj.weight": "model-00033-of-00062.safetensors", "model.layers.45.self_attn.o_proj.weight": "model-00033-of-00062.safetensors", "model.layers.45.self_attn.q_proj.weight": "model-00033-of-00062.safetensors", "model.layers.45.self_attn.v_proj.weight": "model-00033-of-00062.safetensors", "model.layers.46.input_layernorm.weight": "model-00033-of-00062.safetensors", "model.layers.46.mlp.down_proj.weight": "model-00033-of-00062.safetensors", "model.layers.46.mlp.gate_proj.weight": "model-00034-of-00062.safetensors", "model.layers.46.mlp.up_proj.weight": "model-00034-of-00062.safetensors", "model.layers.46.post_attention_layernorm.weight": "model-00034-of-00062.safetensors", "model.layers.46.self_attn.k_proj.weight": "model-00034-of-00062.safetensors", "model.layers.46.self_attn.o_proj.weight": "model-00034-of-00062.safetensors", "model.layers.46.self_attn.q_proj.weight": "model-00034-of-00062.safetensors", "model.layers.46.self_attn.v_proj.weight": "model-00034-of-00062.safetensors", "model.layers.47.input_layernorm.weight": "model-00034-of-00062.safetensors", "model.layers.47.mlp.down_proj.weight": "model-00034-of-00062.safetensors", "model.layers.47.mlp.gate_proj.weight": "model-00034-of-00062.safetensors", "model.layers.47.mlp.up_proj.weight": "model-00035-of-00062.safetensors", "model.layers.47.post_attention_layernorm.weight": "model-00035-of-00062.safetensors", "model.layers.47.self_attn.k_proj.weight": "model-00035-of-00062.safetensors", "model.layers.47.self_attn.o_proj.weight": "model-00035-of-00062.safetensors", "model.layers.47.self_attn.q_proj.weight": "model-00035-of-00062.safetensors", "model.layers.47.self_attn.v_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.input_layernorm.weight": "model-00035-of-00062.safetensors", "model.layers.48.mlp.down_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.mlp.gate_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.mlp.up_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.post_attention_layernorm.weight": "model-00035-of-00062.safetensors", "model.layers.48.self_attn.k_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.self_attn.o_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.self_attn.q_proj.weight": "model-00035-of-00062.safetensors", "model.layers.48.self_attn.v_proj.weight": "model-00035-of-00062.safetensors", "model.layers.49.input_layernorm.weight": "model-00035-of-00062.safetensors", "model.layers.49.mlp.down_proj.weight": "model-00036-of-00062.safetensors", "model.layers.49.mlp.gate_proj.weight": "model-00036-of-00062.safetensors", "model.layers.49.mlp.up_proj.weight": "model-00036-of-00062.safetensors", "model.layers.49.post_attention_layernorm.weight": "model-00036-of-00062.safetensors", "model.layers.49.self_attn.k_proj.weight": "model-00036-of-00062.safetensors", "model.layers.49.self_attn.o_proj.weight": "model-00036-of-00062.safetensors", "model.layers.49.self_attn.q_proj.weight": "model-00036-of-00062.safetensors", "model.layers.49.self_attn.v_proj.weight": "model-00036-of-00062.safetensors", "model.layers.5.input_layernorm.weight": "model-00036-of-00062.safetensors", "model.layers.5.mlp.down_proj.weight": "model-00036-of-00062.safetensors", "model.layers.5.mlp.gate_proj.weight": "model-00037-of-00062.safetensors", "model.layers.5.mlp.up_proj.weight": "model-00037-of-00062.safetensors", "model.layers.5.post_attention_layernorm.weight": "model-00037-of-00062.safetensors", "model.layers.5.self_attn.k_proj.weight": "model-00037-of-00062.safetensors", "model.layers.5.self_attn.o_proj.weight": "model-00037-of-00062.safetensors", "model.layers.5.self_attn.q_proj.weight": "model-00037-of-00062.safetensors", "model.layers.5.self_attn.v_proj.weight": "model-00037-of-00062.safetensors", "model.layers.50.input_layernorm.weight": "model-00037-of-00062.safetensors", "model.layers.50.mlp.down_proj.weight": "model-00037-of-00062.safetensors", "model.layers.50.mlp.gate_proj.weight": "model-00037-of-00062.safetensors", "model.layers.50.mlp.up_proj.weight": "model-00038-of-00062.safetensors", "model.layers.50.post_attention_layernorm.weight": "model-00038-of-00062.safetensors", "model.layers.50.self_attn.k_proj.weight": "model-00038-of-00062.safetensors", "model.layers.50.self_attn.o_proj.weight": "model-00038-of-00062.safetensors", "model.layers.50.self_attn.q_proj.weight": "model-00038-of-00062.safetensors", "model.layers.50.self_attn.v_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.input_layernorm.weight": "model-00038-of-00062.safetensors", "model.layers.51.mlp.down_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.mlp.gate_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.mlp.up_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.post_attention_layernorm.weight": "model-00038-of-00062.safetensors", "model.layers.51.self_attn.k_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.self_attn.o_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.self_attn.q_proj.weight": "model-00038-of-00062.safetensors", "model.layers.51.self_attn.v_proj.weight": "model-00038-of-00062.safetensors", "model.layers.52.input_layernorm.weight": "model-00038-of-00062.safetensors", "model.layers.52.mlp.down_proj.weight": "model-00039-of-00062.safetensors", "model.layers.52.mlp.gate_proj.weight": "model-00039-of-00062.safetensors", "model.layers.52.mlp.up_proj.weight": "model-00039-of-00062.safetensors", "model.layers.52.post_attention_layernorm.weight": "model-00039-of-00062.safetensors", "model.layers.52.self_attn.k_proj.weight": "model-00039-of-00062.safetensors", "model.layers.52.self_attn.o_proj.weight": "model-00039-of-00062.safetensors", "model.layers.52.self_attn.q_proj.weight": "model-00039-of-00062.safetensors", "model.layers.52.self_attn.v_proj.weight": "model-00039-of-00062.safetensors", "model.layers.53.input_layernorm.weight": "model-00039-of-00062.safetensors", "model.layers.53.mlp.down_proj.weight": "model-00039-of-00062.safetensors", "model.layers.53.mlp.gate_proj.weight": "model-00040-of-00062.safetensors", "model.layers.53.mlp.up_proj.weight": "model-00040-of-00062.safetensors", "model.layers.53.post_attention_layernorm.weight": "model-00040-of-00062.safetensors", "model.layers.53.self_attn.k_proj.weight": "model-00040-of-00062.safetensors", "model.layers.53.self_attn.o_proj.weight": "model-00040-of-00062.safetensors", "model.layers.53.self_attn.q_proj.weight": "model-00040-of-00062.safetensors", "model.layers.53.self_attn.v_proj.weight": "model-00040-of-00062.safetensors", "model.layers.54.input_layernorm.weight": "model-00040-of-00062.safetensors", "model.layers.54.mlp.down_proj.weight": "model-00040-of-00062.safetensors", "model.layers.54.mlp.gate_proj.weight": "model-00040-of-00062.safetensors", "model.layers.54.mlp.up_proj.weight": "model-00041-of-00062.safetensors", "model.layers.54.post_attention_layernorm.weight": "model-00041-of-00062.safetensors", "model.layers.54.self_attn.k_proj.weight": "model-00041-of-00062.safetensors", "model.layers.54.self_attn.o_proj.weight": "model-00041-of-00062.safetensors", "model.layers.54.self_attn.q_proj.weight": "model-00041-of-00062.safetensors", "model.layers.54.self_attn.v_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.input_layernorm.weight": "model-00041-of-00062.safetensors", "model.layers.55.mlp.down_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.mlp.gate_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.mlp.up_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.post_attention_layernorm.weight": "model-00041-of-00062.safetensors", "model.layers.55.self_attn.k_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.self_attn.o_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.self_attn.q_proj.weight": "model-00041-of-00062.safetensors", "model.layers.55.self_attn.v_proj.weight": "model-00041-of-00062.safetensors", "model.layers.56.input_layernorm.weight": "model-00041-of-00062.safetensors", "model.layers.56.mlp.down_proj.weight": "model-00042-of-00062.safetensors", "model.layers.56.mlp.gate_proj.weight": "model-00042-of-00062.safetensors", "model.layers.56.mlp.up_proj.weight": "model-00042-of-00062.safetensors", "model.layers.56.post_attention_layernorm.weight": "model-00042-of-00062.safetensors", "model.layers.56.self_attn.k_proj.weight": "model-00042-of-00062.safetensors", "model.layers.56.self_attn.o_proj.weight": "model-00042-of-00062.safetensors", "model.layers.56.self_attn.q_proj.weight": "model-00042-of-00062.safetensors", "model.layers.56.self_attn.v_proj.weight": "model-00042-of-00062.safetensors", "model.layers.57.input_layernorm.weight": "model-00042-of-00062.safetensors", "model.layers.57.mlp.down_proj.weight": "model-00042-of-00062.safetensors", "model.layers.57.mlp.gate_proj.weight": "model-00043-of-00062.safetensors", "model.layers.57.mlp.up_proj.weight": "model-00043-of-00062.safetensors", "model.layers.57.post_attention_layernorm.weight": "model-00043-of-00062.safetensors", "model.layers.57.self_attn.k_proj.weight": "model-00043-of-00062.safetensors", "model.layers.57.self_attn.o_proj.weight": "model-00043-of-00062.safetensors", "model.layers.57.self_attn.q_proj.weight": "model-00043-of-00062.safetensors", "model.layers.57.self_attn.v_proj.weight": "model-00043-of-00062.safetensors", "model.layers.58.input_layernorm.weight": "model-00043-of-00062.safetensors", "model.layers.58.mlp.down_proj.weight": "model-00043-of-00062.safetensors", "model.layers.58.mlp.gate_proj.weight": "model-00043-of-00062.safetensors", "model.layers.58.mlp.up_proj.weight": "model-00044-of-00062.safetensors", "model.layers.58.post_attention_layernorm.weight": "model-00044-of-00062.safetensors", "model.layers.58.self_attn.k_proj.weight": "model-00044-of-00062.safetensors", "model.layers.58.self_attn.o_proj.weight": "model-00044-of-00062.safetensors", "model.layers.58.self_attn.q_proj.weight": "model-00044-of-00062.safetensors", "model.layers.58.self_attn.v_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.input_layernorm.weight": "model-00044-of-00062.safetensors", "model.layers.59.mlp.down_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.mlp.gate_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.mlp.up_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.post_attention_layernorm.weight": "model-00044-of-00062.safetensors", "model.layers.59.self_attn.k_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.self_attn.o_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.self_attn.q_proj.weight": "model-00044-of-00062.safetensors", "model.layers.59.self_attn.v_proj.weight": "model-00044-of-00062.safetensors", "model.layers.6.input_layernorm.weight": "model-00044-of-00062.safetensors", "model.layers.6.mlp.down_proj.weight": "model-00045-of-00062.safetensors", "model.layers.6.mlp.gate_proj.weight": "model-00045-of-00062.safetensors", "model.layers.6.mlp.up_proj.weight": "model-00045-of-00062.safetensors", "model.layers.6.post_attention_layernorm.weight": "model-00045-of-00062.safetensors", "model.layers.6.self_attn.k_proj.weight": "model-00045-of-00062.safetensors", "model.layers.6.self_attn.o_proj.weight": "model-00045-of-00062.safetensors", "model.layers.6.self_attn.q_proj.weight": "model-00045-of-00062.safetensors", "model.layers.6.self_attn.v_proj.weight": "model-00045-of-00062.safetensors", "model.layers.60.input_layernorm.weight": "model-00045-of-00062.safetensors", "model.layers.60.mlp.down_proj.weight": "model-00045-of-00062.safetensors", "model.layers.60.mlp.gate_proj.weight": "model-00046-of-00062.safetensors", "model.layers.60.mlp.up_proj.weight": "model-00046-of-00062.safetensors", "model.layers.60.post_attention_layernorm.weight": "model-00046-of-00062.safetensors", "model.layers.60.self_attn.k_proj.weight": "model-00046-of-00062.safetensors", "model.layers.60.self_attn.o_proj.weight": "model-00046-of-00062.safetensors", "model.layers.60.self_attn.q_proj.weight": "model-00046-of-00062.safetensors", "model.layers.60.self_attn.v_proj.weight": "model-00046-of-00062.safetensors", "model.layers.61.input_layernorm.weight": "model-00046-of-00062.safetensors", "model.layers.61.mlp.down_proj.weight": "model-00046-of-00062.safetensors", "model.layers.61.mlp.gate_proj.weight": "model-00046-of-00062.safetensors", "model.layers.61.mlp.up_proj.weight": "model-00047-of-00062.safetensors", "model.layers.61.post_attention_layernorm.weight": "model-00047-of-00062.safetensors", "model.layers.61.self_attn.k_proj.weight": "model-00047-of-00062.safetensors", "model.layers.61.self_attn.o_proj.weight": "model-00047-of-00062.safetensors", "model.layers.61.self_attn.q_proj.weight": "model-00047-of-00062.safetensors", "model.layers.61.self_attn.v_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.input_layernorm.weight": "model-00047-of-00062.safetensors", "model.layers.62.mlp.down_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.mlp.gate_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.mlp.up_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.post_attention_layernorm.weight": "model-00047-of-00062.safetensors", "model.layers.62.self_attn.k_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.self_attn.o_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.self_attn.q_proj.weight": "model-00047-of-00062.safetensors", "model.layers.62.self_attn.v_proj.weight": "model-00047-of-00062.safetensors", "model.layers.63.input_layernorm.weight": "model-00047-of-00062.safetensors", "model.layers.63.mlp.down_proj.weight": "model-00048-of-00062.safetensors", "model.layers.63.mlp.gate_proj.weight": "model-00048-of-00062.safetensors", "model.layers.63.mlp.up_proj.weight": "model-00048-of-00062.safetensors", "model.layers.63.post_attention_layernorm.weight": "model-00048-of-00062.safetensors", "model.layers.63.self_attn.k_proj.weight": "model-00048-of-00062.safetensors", "model.layers.63.self_attn.o_proj.weight": "model-00048-of-00062.safetensors", "model.layers.63.self_attn.q_proj.weight": "model-00048-of-00062.safetensors", "model.layers.63.self_attn.v_proj.weight": "model-00048-of-00062.safetensors", "model.layers.64.input_layernorm.weight": "model-00048-of-00062.safetensors", "model.layers.64.mlp.down_proj.weight": "model-00048-of-00062.safetensors", "model.layers.64.mlp.gate_proj.weight": "model-00049-of-00062.safetensors", "model.layers.64.mlp.up_proj.weight": "model-00049-of-00062.safetensors", "model.layers.64.post_attention_layernorm.weight": "model-00049-of-00062.safetensors", "model.layers.64.self_attn.k_proj.weight": "model-00049-of-00062.safetensors", "model.layers.64.self_attn.o_proj.weight": "model-00049-of-00062.safetensors", "model.layers.64.self_attn.q_proj.weight": "model-00049-of-00062.safetensors", "model.layers.64.self_attn.v_proj.weight": "model-00049-of-00062.safetensors", "model.layers.65.input_layernorm.weight": "model-00049-of-00062.safetensors", "model.layers.65.mlp.down_proj.weight": "model-00049-of-00062.safetensors", "model.layers.65.mlp.gate_proj.weight": "model-00049-of-00062.safetensors", "model.layers.65.mlp.up_proj.weight": "model-00050-of-00062.safetensors", "model.layers.65.post_attention_layernorm.weight": "model-00050-of-00062.safetensors", "model.layers.65.self_attn.k_proj.weight": "model-00050-of-00062.safetensors", "model.layers.65.self_attn.o_proj.weight": "model-00050-of-00062.safetensors", "model.layers.65.self_attn.q_proj.weight": "model-00050-of-00062.safetensors", "model.layers.65.self_attn.v_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.input_layernorm.weight": "model-00050-of-00062.safetensors", "model.layers.66.mlp.down_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.mlp.gate_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.mlp.up_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.post_attention_layernorm.weight": "model-00050-of-00062.safetensors", "model.layers.66.self_attn.k_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.self_attn.o_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.self_attn.q_proj.weight": "model-00050-of-00062.safetensors", "model.layers.66.self_attn.v_proj.weight": "model-00050-of-00062.safetensors", "model.layers.67.input_layernorm.weight": "model-00050-of-00062.safetensors", "model.layers.67.mlp.down_proj.weight": "model-00051-of-00062.safetensors", "model.layers.67.mlp.gate_proj.weight": "model-00051-of-00062.safetensors", "model.layers.67.mlp.up_proj.weight": "model-00051-of-00062.safetensors", "model.layers.67.post_attention_layernorm.weight": "model-00051-of-00062.safetensors", "model.layers.67.self_attn.k_proj.weight": "model-00051-of-00062.safetensors", "model.layers.67.self_attn.o_proj.weight": "model-00051-of-00062.safetensors", "model.layers.67.self_attn.q_proj.weight": "model-00051-of-00062.safetensors", "model.layers.67.self_attn.v_proj.weight": "model-00051-of-00062.safetensors", "model.layers.68.input_layernorm.weight": "model-00051-of-00062.safetensors", "model.layers.68.mlp.down_proj.weight": "model-00051-of-00062.safetensors", "model.layers.68.mlp.gate_proj.weight": "model-00052-of-00062.safetensors", "model.layers.68.mlp.up_proj.weight": "model-00052-of-00062.safetensors", "model.layers.68.post_attention_layernorm.weight": "model-00052-of-00062.safetensors", "model.layers.68.self_attn.k_proj.weight": "model-00052-of-00062.safetensors", "model.layers.68.self_attn.o_proj.weight": "model-00052-of-00062.safetensors", "model.layers.68.self_attn.q_proj.weight": "model-00052-of-00062.safetensors", "model.layers.68.self_attn.v_proj.weight": "model-00052-of-00062.safetensors", "model.layers.69.input_layernorm.weight": "model-00052-of-00062.safetensors", "model.layers.69.mlp.down_proj.weight": "model-00052-of-00062.safetensors", "model.layers.69.mlp.gate_proj.weight": "model-00052-of-00062.safetensors", "model.layers.69.mlp.up_proj.weight": "model-00053-of-00062.safetensors", "model.layers.69.post_attention_layernorm.weight": "model-00053-of-00062.safetensors", "model.layers.69.self_attn.k_proj.weight": "model-00053-of-00062.safetensors", "model.layers.69.self_attn.o_proj.weight": "model-00053-of-00062.safetensors", "model.layers.69.self_attn.q_proj.weight": "model-00053-of-00062.safetensors", "model.layers.69.self_attn.v_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.input_layernorm.weight": "model-00053-of-00062.safetensors", "model.layers.7.mlp.down_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.mlp.gate_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.mlp.up_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.post_attention_layernorm.weight": "model-00053-of-00062.safetensors", "model.layers.7.self_attn.k_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.self_attn.o_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.self_attn.q_proj.weight": "model-00053-of-00062.safetensors", "model.layers.7.self_attn.v_proj.weight": "model-00053-of-00062.safetensors", "model.layers.70.input_layernorm.weight": "model-00053-of-00062.safetensors", "model.layers.70.mlp.down_proj.weight": "model-00054-of-00062.safetensors", "model.layers.70.mlp.gate_proj.weight": "model-00054-of-00062.safetensors", "model.layers.70.mlp.up_proj.weight": "model-00054-of-00062.safetensors", "model.layers.70.post_attention_layernorm.weight": "model-00054-of-00062.safetensors", "model.layers.70.self_attn.k_proj.weight": "model-00054-of-00062.safetensors", "model.layers.70.self_attn.o_proj.weight": "model-00054-of-00062.safetensors", "model.layers.70.self_attn.q_proj.weight": "model-00054-of-00062.safetensors", "model.layers.70.self_attn.v_proj.weight": "model-00054-of-00062.safetensors", "model.layers.71.input_layernorm.weight": "model-00054-of-00062.safetensors", "model.layers.71.mlp.down_proj.weight": "model-00054-of-00062.safetensors", "model.layers.71.mlp.gate_proj.weight": "model-00055-of-00062.safetensors", "model.layers.71.mlp.up_proj.weight": "model-00055-of-00062.safetensors", "model.layers.71.post_attention_layernorm.weight": "model-00055-of-00062.safetensors", "model.layers.71.self_attn.k_proj.weight": "model-00055-of-00062.safetensors", "model.layers.71.self_attn.o_proj.weight": "model-00055-of-00062.safetensors", "model.layers.71.self_attn.q_proj.weight": "model-00055-of-00062.safetensors", "model.layers.71.self_attn.v_proj.weight": "model-00055-of-00062.safetensors", "model.layers.72.input_layernorm.weight": "model-00055-of-00062.safetensors", "model.layers.72.mlp.down_proj.weight": "model-00055-of-00062.safetensors", "model.layers.72.mlp.gate_proj.weight": "model-00055-of-00062.safetensors", "model.layers.72.mlp.up_proj.weight": "model-00056-of-00062.safetensors", "model.layers.72.post_attention_layernorm.weight": "model-00056-of-00062.safetensors", "model.layers.72.self_attn.k_proj.weight": "model-00056-of-00062.safetensors", "model.layers.72.self_attn.o_proj.weight": "model-00056-of-00062.safetensors", "model.layers.72.self_attn.q_proj.weight": "model-00056-of-00062.safetensors", "model.layers.72.self_attn.v_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.input_layernorm.weight": "model-00056-of-00062.safetensors", "model.layers.73.mlp.down_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.mlp.gate_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.mlp.up_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.post_attention_layernorm.weight": "model-00056-of-00062.safetensors", "model.layers.73.self_attn.k_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.self_attn.o_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.self_attn.q_proj.weight": "model-00056-of-00062.safetensors", "model.layers.73.self_attn.v_proj.weight": "model-00056-of-00062.safetensors", "model.layers.74.input_layernorm.weight": "model-00056-of-00062.safetensors", "model.layers.74.mlp.down_proj.weight": "model-00057-of-00062.safetensors", "model.layers.74.mlp.gate_proj.weight": "model-00057-of-00062.safetensors", "model.layers.74.mlp.up_proj.weight": "model-00057-of-00062.safetensors", "model.layers.74.post_attention_layernorm.weight": "model-00057-of-00062.safetensors", "model.layers.74.self_attn.k_proj.weight": "model-00057-of-00062.safetensors", "model.layers.74.self_attn.o_proj.weight": "model-00057-of-00062.safetensors", "model.layers.74.self_attn.q_proj.weight": "model-00057-of-00062.safetensors", "model.layers.74.self_attn.v_proj.weight": "model-00057-of-00062.safetensors", "model.layers.75.input_layernorm.weight": "model-00057-of-00062.safetensors", "model.layers.75.mlp.down_proj.weight": "model-00057-of-00062.safetensors", "model.layers.75.mlp.gate_proj.weight": "model-00058-of-00062.safetensors", "model.layers.75.mlp.up_proj.weight": "model-00058-of-00062.safetensors", "model.layers.75.post_attention_layernorm.weight": "model-00058-of-00062.safetensors", "model.layers.75.self_attn.k_proj.weight": "model-00058-of-00062.safetensors", "model.layers.75.self_attn.o_proj.weight": "model-00058-of-00062.safetensors", "model.layers.75.self_attn.q_proj.weight": "model-00058-of-00062.safetensors", "model.layers.75.self_attn.v_proj.weight": "model-00058-of-00062.safetensors", "model.layers.76.input_layernorm.weight": "model-00058-of-00062.safetensors", "model.layers.76.mlp.down_proj.weight": "model-00058-of-00062.safetensors", "model.layers.76.mlp.gate_proj.weight": "model-00058-of-00062.safetensors", "model.layers.76.mlp.up_proj.weight": "model-00059-of-00062.safetensors", "model.layers.76.post_attention_layernorm.weight": "model-00059-of-00062.safetensors", "model.layers.76.self_attn.k_proj.weight": "model-00059-of-00062.safetensors", "model.layers.76.self_attn.o_proj.weight": "model-00059-of-00062.safetensors", "model.layers.76.self_attn.q_proj.weight": "model-00059-of-00062.safetensors", "model.layers.76.self_attn.v_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.input_layernorm.weight": "model-00059-of-00062.safetensors", "model.layers.77.mlp.down_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.mlp.gate_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.mlp.up_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.post_attention_layernorm.weight": "model-00059-of-00062.safetensors", "model.layers.77.self_attn.k_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.self_attn.o_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.self_attn.q_proj.weight": "model-00059-of-00062.safetensors", "model.layers.77.self_attn.v_proj.weight": "model-00059-of-00062.safetensors", "model.layers.78.input_layernorm.weight": "model-00059-of-00062.safetensors", "model.layers.78.mlp.down_proj.weight": "model-00060-of-00062.safetensors", "model.layers.78.mlp.gate_proj.weight": "model-00060-of-00062.safetensors", "model.layers.78.mlp.up_proj.weight": "model-00060-of-00062.safetensors", "model.layers.78.post_attention_layernorm.weight": "model-00060-of-00062.safetensors", "model.layers.78.self_attn.k_proj.weight": "model-00060-of-00062.safetensors", "model.layers.78.self_attn.o_proj.weight": "model-00060-of-00062.safetensors", "model.layers.78.self_attn.q_proj.weight": "model-00060-of-00062.safetensors", "model.layers.78.self_attn.v_proj.weight": "model-00060-of-00062.safetensors", "model.layers.79.input_layernorm.weight": "model-00060-of-00062.safetensors", "model.layers.79.mlp.down_proj.weight": "model-00060-of-00062.safetensors", "model.layers.79.mlp.gate_proj.weight": "model-00061-of-00062.safetensors", "model.layers.79.mlp.up_proj.weight": "model-00061-of-00062.safetensors", "model.layers.79.post_attention_layernorm.weight": "model-00061-of-00062.safetensors", "model.layers.79.self_attn.k_proj.weight": "model-00061-of-00062.safetensors", "model.layers.79.self_attn.o_proj.weight": "model-00061-of-00062.safetensors", "model.layers.79.self_attn.q_proj.weight": "model-00061-of-00062.safetensors", "model.layers.79.self_attn.v_proj.weight": "model-00061-of-00062.safetensors", "model.layers.8.input_layernorm.weight": "model-00061-of-00062.safetensors", "model.layers.8.mlp.down_proj.weight": "model-00061-of-00062.safetensors", "model.layers.8.mlp.gate_proj.weight": "model-00061-of-00062.safetensors", "model.layers.8.mlp.up_proj.weight": "model-00062-of-00062.safetensors", "model.layers.8.post_attention_layernorm.weight": "model-00062-of-00062.safetensors", "model.layers.8.self_attn.k_proj.weight": "model-00062-of-00062.safetensors", "model.layers.8.self_attn.o_proj.weight": "model-00062-of-00062.safetensors", "model.layers.8.self_attn.q_proj.weight": "model-00062-of-00062.safetensors", "model.layers.8.self_attn.v_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.input_layernorm.weight": "model-00062-of-00062.safetensors", "model.layers.9.mlp.down_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.mlp.gate_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.mlp.up_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.post_attention_layernorm.weight": "model-00062-of-00062.safetensors", "model.layers.9.self_attn.k_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.self_attn.o_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.self_attn.q_proj.weight": "model-00062-of-00062.safetensors", "model.layers.9.self_attn.v_proj.weight": "model-00062-of-00062.safetensors", "model.norm.weight": "model-00062-of-00062.safetensors"}}

.ipynb_checkpoints/special_tokens_map-checkpoint.json

@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

README.md

@@ -1,3 +1,41 @@
----
-license: llama3.1
----
+---
+base_model: []
+library_name: transformers
+tags:
+- mergekit
+- merge
+
+---
+# new-dawn-ultra-llama3-70b-32K-v1.0
+
+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 SLERP merge method.
+
+### Models Merged
+
+The following models were included in the merge:
+* ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K
+* ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.18-32K
+
+### Configuration
+
+The following YAML configuration was used to produce this model:
+
+```yaml
+base_model: ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K
+dtype: float32
+merge_method: slerp
+parameters:
+  t:
+  - value: 0.5
+slices:
+- sources:
+  - layer_range: [0, 80]
+    model: ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K
+  - layer_range: [0, 80]
+    model: ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.18-32K
+```

config.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "hidden_act": "silu",
+  "hidden_size": 8192,
+  "initializer_range": 0.02,
+  "intermediate_size": 28672,
+  "max_position_embeddings": 32768,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 64,
+  "num_hidden_layers": 80,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 3000000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.44.2",
+  "use_cache": true,
+  "use_flash_attention_2": true,
+  "vocab_size": 128256
+}

mergekit_config.yml

@@ -0,0 +1,12 @@
+base_model: ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K
+dtype: float32
+merge_method: slerp
+parameters:
+  t:
+  - value: 0.5
+slices:
+- sources:
+  - layer_range: [0, 80]
+    model: ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.16-32K
+  - layer_range: [0, 80]
+    model: ../mergekit-output/new-dawn-ultra-llama3-70b-32K-v1.0.yaml/new-dawn-ultra-llama3-70b-v0.18-32K

model-00001-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0935b2dd453ea6e532fc676eaf4e60ebab8554e236d8dc789dcee1f7728823f2
+size 4202692736

model-00002-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f68f0e27f4433df66c798051c60eaf402e0c73664c36793de11f34988ffceb1e
+size 4202725624

model-00003-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc56d5b8530b057e7f4ebacc5db45ada422ba3d600c2092373cb60291627441d
+size 4362142872

model-00004-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f57ad0a90761260656f60773ba405cc66f1414ea15e39df0f29838852fee0cd
+size 4362142880

model-00005-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37eb30730167dd219ce29a4f2f2ee25a35748ceb68c75e16d62612aee75f116c
+size 4966188888

model-00006-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91e6fe685942c6514b7647db69fa1e7c1c4e18e496b85a2afdcfe4599d551d7a
+size 4362142888

model-00007-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea5e8b555fc34d35d0fd88dc00a2aac5cba19a3a25fdf1729de053f58c1bbb8b
+size 4362142888

model-00008-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cae01da9f2f26b7818e2a87c62157efb6016153783accddbac0a0d3ddafe8d87
+size 4966188888

model-00009-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6de333940b5335e0a820c042a07557e3546cc3b63d9b4dd3d166d88a69ba0461
+size 4362142888

model-00010-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ad97134b5d632ee9933a7c3a6ab502c5b4c0d53fd3d6f443acf052df341c8e7
+size 4362142888

model-00011-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbae605be0b1227cd3ac263025dac8e8698b3a788ce919230742e76ab6186243
+size 4966188888

model-00012-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e809f4868f8b573d53f5217ebb7ef8c5533c9db99106f6d1db1850fe6f735646
+size 4362142880

model-00013-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c32b139d84d18ef2bdb7b639e97c7e2bdf0e73c1119650519a38b191f6d52329
+size 4362142888

model-00014-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9481f0f6650f646dce7c39abcacdd703a2a96b726f5a21ba6ecb01a0a751642
+size 4966188888

model-00015-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:279e822293a08429c64ed4b501d617bb5c6ca63a6464bfb6d2a7fe15612ea500
+size 4362142888

model-00016-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27528eb34d8fe260f991e9e9b9e0591335025d20bf618447cf08b7ae692a0353
+size 4362142888

model-00017-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa7fc80334af20cb23094f4e16d92c9c900d4eedb6c60d5ed6150c087ba19ca4
+size 4966188888

model-00018-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d4a97063a4448848309cbd0629027c2c4cb5551c58fd5d3d26ec2d046814103
+size 4362142888

model-00019-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c30c47328aa19ba546d05e4fc93a017436632cff962b960987ac7db529eb1df1
+size 4362142888

model-00020-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdc9139a86a70512ba566e50644445ca148db4b699e0507cf1c3611dcbdc5d25
+size 4966188880

model-00021-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fdbc837573323ba4a63deaef5154cbb23dcca1d4f7b78688a52190b2c9641b4
+size 4362142888

model-00022-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7a603fe6ecc2587633a8fc0231c9b70698d47823e914648f6f7515273000156
+size 4362142888

model-00023-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0c981f7729df18e2411ad3fb0894ee158cc6d2d70210d9f2b243d1a24f4696d
+size 4966188888

model-00024-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc9ca333ccd63c5a60291899cf89a726525589cdbb56d01a8f417466c990929b
+size 4362142888

model-00025-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0086ec0dbb17d6913ab21a500adc8fce7f12025abbeaf29f420f3dc68b7cda37
+size 4362142888

model-00026-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:706a5dd5bf0e4429f1a34ce7d30ec36b6efdcd00ebe6b1681a35a7b7423f2541
+size 4966188888

model-00027-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13838c80810c6d9903f2ec103f897b8eb321f33e4e8e86e0abf5724d5fd4e411
+size 4362142888

model-00028-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:829be77e35cfb6d6e17f9afb9f4b1835c700246ba85efa712c9c81915cd1e487
+size 4362142880

model-00029-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0461462700ec84d219a1f9240c3b0ea6906df4ac6e19c1fcd89af00fcb9cc770
+size 4966188888

model-00030-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb1ebfca3f9165d8d48ada86c0f068220359dee40f178ad17d539da352844751
+size 4362142888

model-00031-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9a74292a9a0b1a7acd55d16285460a6ab40f03c3a17a7fcd54277bd62705973
+size 4362142888

model-00032-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c56b425641d3742f125d8f6482c004f3101d3fdd57e065525f3d147ae32df5b
+size 4966188888

model-00033-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73521805e62beabc48b5b8eb27712a31eba028cc3ba4c4d8bba437b265efa742
+size 4362142888

model-00034-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8a30e493250d9136402ca26bda421428c6dd25ead25db3eb0a9e9e4dc4a6ee5
+size 4362142888

model-00035-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71236c6483b2251f7caa5e3fd692cb5ddfa9c845e210539094e6d9d4116f73c3
+size 4966188888

model-00036-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cea5d7032847c0c42703a2c82a7ffb698df90ff5614dacd4db773b057fba959a
+size 4362142880

model-00037-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:095f3c4a62d1fd83444dd77a23ed54b4aaa0a39d64e4e33d38880b7a31daefb5
+size 4362142880

model-00038-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c5e786ca35f37c16db6103ade5064a179e92547258afd73e1fa88d8bad6e1a7
+size 4966188888

model-00039-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f779c50d4ebce30c570c845e0eb01ec98002c9eab764f18eafddf5422bce0bc3
+size 4362142888

model-00040-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba31bf517c4133a646fcd50a34b206c6b4ba6ffb66345d1f27404bda67adf544
+size 4362142888

model-00041-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c347e9ef7c225b9814c0b5266410762a40e7d53563759baf13939d8f7671813
+size 4966188888

model-00042-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1509e92d7c627388faed9803d1e28b96dd72a5e64f9356bebf480dd4a74e6463
+size 4362142888

model-00043-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73ffc443f52e919aa293e2cb405f7016257223435ea00ddcfc35593425c865b8
+size 4362142888

model-00044-of-00062.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cdb900b5a18e76138f78c2d16a8ef81f80e955a62cf2df5f8a50aeaa455581b
+size 4966188888