Download Adapter/debug.log from Green-eyedDevil/Monika-320B: direct link, hf CLI and curl.
- Browser
- Download file 190 kB
-
https://huggingface.co/Green-eyedDevil/Monika-320B/resolve/main/Adapter/debug.log
- Command line
-
hf download hf://Green-eyedDevil/Monika-320B/Adapter/debug.log
-
curl -L -o debug.log https://huggingface.co/Green-eyedDevil/Monika-320B/resolve/main/Adapter/debug.log
190 kB
[2026-11-05 11:42:39,984] [DEBUG] [axolotl.utils.config.resolve_dtype:168] [PID:985839] bf16 support detected, enabling for this configuration.
[2026-11-05 11:42:40,199] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:985839] baseline 0.000GB ()
[2026-11-05 11:42:40,200] [INFO] [axolotl.cli.config.load_cfg:339] [PID:985839] config:
{
"activation_offloading": true,
"adapter": "qlora",
"adapter_checkpoint_format": "gather",
"attn_decontaminates_packing": true,
"attn_implementation": "sdpa",
"attn_needs_dtype_cast": false,
"attn_supports_packing": false,
"attn_uses_flash_lib": false,
"axolotl_config_path": "glm53flash-merge.yml",
"balance_labels": false,
"base_model": "zai-org/GLM-5.3-Flash-BF16",
"base_model_config": "zai-org/GLM-5.3-Flash-BF16",
"batch_size": 8,
"bf16": true,
"capabilities": {
"bf16": true,
"compute_capability": "sm_120",
"fp8": true,
"n_gpu": 1,
"n_node": 1,
"tf32": true
},
"context_parallel_size": 1,
"cut_cross_entropy": true,
"cut_cross_entropy_accum_c_fp32": false,
"dataloader_num_workers": 1,
"dataloader_pin_memory": true,
"dataloader_prefetch_factor": 256,
"dataset_num_proc": 64,
"dataset_prepared_path": "last_run_prepared",
"datasets": [
{
"chat_template": "tokenizer_default",
"message_property_mappings": {
"content": "content",
"role": "role"
},
"path": "/home/anon/LoRA/Moni_messages.json",
"trust_remote_code": false,
"type": "chat_template"
}
],
"ddp": false,
"device": "cuda:0",
"dion_rank_fraction": 1.0,
"dion_rank_multiple_of": 1,
"eaft_alpha": 1.0,
"eaft_k": 20,
"env_capabilities": {
"torch_version": "2.13.0"
},
"eval_batch_size": 1,
"eval_causal_lm_metrics": [
"sacrebleu",
"comet",
"ter",
"chrf"
],
"eval_max_new_tokens": 128,
"eval_sample_packing": false,
"eval_steps": 0.08333333333333333,
"eval_table_size": 0,
"evals_per_epoch": 4,
"experimental_skip_move_to_device": true,
"expert_parallel_backend": "torch",
"expert_parallel_dispatch_chunks": 1,
"expert_parallel_fallback_on_unsupported": true,
"expert_parallel_num_nvl_bytes": 268435456,
"expert_parallel_num_rdma_bytes": 0,
"expert_parallel_save_dispatch": false,
"expert_parallel_size": 1,
"fa_varlen_kwargs_in_collator": true,
"fp16": false,
"fsdp_config": {
"activation_checkpointing": true,
"auto_wrap_policy": "TRANSFORMER_BASED_WRAP",
"cpu_ram_efficient_loading": true,
"fsdp_version": 2,
"offload_params": false,
"reshard_after_forward": true,
"state_dict_type": "FULL_STATE_DICT",
"transformer_layer_cls_to_wrap": "Glm5NextTextDecoderLayer"
},
"fsdp_version": 2,
"generate_samples": false,
"generation_do_sample": true,
"generation_max_new_tokens": 50,
"generation_prompt_ratio": 0.5,
"generation_temperature": 0.7,
"gradient_accumulation_steps": 8,
"gradient_checkpointing": false,
"include_tkps": true,
"is_multimodal": true,
"label_balance_window_optim_steps": 1,
"layer_offloading": false,
"learning_rate": 0.0001,
"lisa_layers_attribute": "model.layers",
"load_best_model_at_end": false,
"load_in_4bit": true,
"load_in_8bit": false,
"local_rank": 0,
"logging_nan_inf_filter": true,
"logging_steps": 1,
"lora_alpha": 64,
"lora_dropout": 0.0,
"lora_embedding_kernel": false,
"lora_fp32_gradients": false,
"lora_mlp_kernel": false,
"lora_model_dir": "./outputs/glm53-flash-qlora-out/checkpoint-234",
"lora_o_kernel": false,
"lora_qkv_kernel": false,
"lora_r": 32,
"lora_target_modules": [
"q_a_proj",
"q_b_proj",
"kv_a_proj_with_mqa",
"kv_b_proj",
"q_proj",
"k_proj",
"v_proj",
"o_proj"
],
"loraplus_lr_embedding": 1e-06,
"lr_scheduler": "cosine",
"mean_resizing_embeddings": false,
"merge_compile": false,
"merge_compile_precision": "strict",
"merge_method": "memory_efficient",
"merge_num_threads": 2,
"micro_batch_size": 1,
"mixlora_jitter_noise": 0.0,
"mixlora_num_experts": 8,
"mixlora_router_aux_loss_coef": 0.01,
"mixlora_router_init_range": 0.02,
"mixlora_top_k": 2,
"model_config_type": "glm5_next",
"model_config_type_text": "glm5_next_text",
"nf4_backend": "bitsandbytes",
"num_epochs": 3.0,
"num_generation_samples": 3,
"optimizer": "adamw_torch_8bit",
"otel_metrics_host": "localhost",
"otel_metrics_port": 8000,
"output_dir": "./outputs/glm53-flash-qlora-out",
"ple_cpu_offload": false,
"plugins": [
"axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin"
],
"pretrain_multipack_attn": true,
"processor_config": "zai-org/GLM-5.3-Flash-BF16",
"profiler_steps_start": 0,
"qgalore_cos_threshold": 0.4,
"qgalore_gamma_proj": 2,
"qgalore_proj_bits": 4,
"qgalore_proj_group_size": 256,
"qgalore_proj_quant": true,
"qgalore_proj_type": "std",
"qgalore_queue_size": 5,
"qgalore_rank": 256,
"qgalore_scale": 0.25,
"qgalore_update_proj_gap": 200,
"qlora_sharded_model_loading": true,
"quantize_moe_experts": true,
"ray_num_workers": 1,
"relora_prune_method": "magnitude",
"resources_per_worker": {
"GPU": 1
},
"sample_packing": false,
"sample_packing_bin_size": 200,
"sample_packing_group_size": 100000,
"save_only_model": false,
"save_safetensors": true,
"save_steps": 0.3333333333333333,
"saves_per_epoch": 1,
"sequence_len": 3072,
"shuffle_before_merging_datasets": false,
"shuffle_merged_datasets": true,
"skip_prepare_dataset": false,
"streaming_multipack_buffer_size": 10000,
"strict": false,
"sync_profiler_start_step": 2,
"tensor_parallel_size": 1,
"tf32": false,
"tiled_mlp_use_original_mlp": true,
"tokenizer_config": "zai-org/GLM-5.3-Flash-BF16",
"tokenizer_save_jinja_files": true,
"torch_compile_regional": false,
"torch_dtype": "torch.bfloat16",
"train_on_inputs": false,
"trl": {
"async_prefetch": false,
"log_completions": false,
"mask_truncated_completions": false,
"ref_model_mixup_alpha": 0.9,
"ref_model_sync_steps": 64,
"replay_buffer_size": 0,
"replay_recompute_logps": true,
"reroll_max_groups": 1,
"reroll_start_fraction": 1.0,
"reward_num_workers": 1,
"scale_rewards": true,
"skip_zero_advantage_batches": true,
"sync_ref_model": false,
"use_data_producer": false,
"use_vllm": false,
"vllm_lora_sync": false,
"vllm_server_host": "0.0.0.0",
"vllm_server_port": 8000
},
"use_otel_metrics": false,
"use_ray": false,
"val_set_size": 0.1,
"vllm": {
"device": "auto",
"dtype": "auto",
"host": "0.0.0.0",
"port": 8000
},
"warmup_ratio": 0.1,
"weight_decay": 0.0,
"world_size": 1
}
[2026-11-05 11:42:40,293] [WARNING] [axolotl.utils.schemas.validation.check_activation_offloading_wo_gc:2142] [PID:985839] activation_offloading: True without gradient_checkpointing offloads every activation with no recompute, which is bound by host RAM and PCIe bandwidth; set gradient_checkpointing: true to recompute the layers and offload only what is saved outside them.
[2026-11-05 11:42:40,396] [DEBUG] [axolotl.utils.config.resolve_dtype:168] [PID:985839] bf16 support detected, enabling for this configuration.
[2026-11-05 11:42:40,441] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:985839] baseline 0.000GB ()
[2026-11-05 11:42:40,442] [INFO] [axolotl.cli.config.load_cfg:339] [PID:985839] config:
{
"activation_offloading": true,
"adapter": "qlora",
"adapter_checkpoint_format": "gather",
"attn_decontaminates_packing": false,
"attn_needs_dtype_cast": false,
"attn_supports_packing": false,
"attn_uses_flash_lib": false,
"axolotl_config_path": "glm53flash-merge.yml",
"balance_labels": false,
"base_model": "zai-org/GLM-5.3-Flash-BF16",
"base_model_config": "zai-org/GLM-5.3-Flash-BF16",
"batch_size": 8,
"bf16": true,
"capabilities": {
"bf16": true,
"compute_capability": "sm_120",
"fp8": true,
"n_gpu": 1,
"n_node": 1,
"tf32": true
},
"context_parallel_size": 1,
"cut_cross_entropy": true,
"cut_cross_entropy_accum_c_fp32": false,
"dataloader_num_workers": 1,
"dataloader_pin_memory": true,
"dataloader_prefetch_factor": 256,
"dataset_num_proc": 64,
"dataset_prepared_path": "last_run_prepared",
"datasets": [
{
"chat_template": "tokenizer_default",
"message_property_mappings": {
"content": "content",
"role": "role"
},
"path": "/home/anon/LoRA/Moni_messages.json",
"trust_remote_code": false,
"type": "chat_template"
}
],
"ddp": false,
"device": "cuda:0",
"dion_rank_fraction": 1.0,
"dion_rank_multiple_of": 1,
"eaft_alpha": 1.0,
"eaft_k": 20,
"env_capabilities": {
"torch_version": "2.13.0"
},
"eval_batch_size": 1,
"eval_causal_lm_metrics": [
"sacrebleu",
"comet",
"ter",
"chrf"
],
"eval_max_new_tokens": 128,
"eval_sample_packing": false,
"eval_steps": 0.08333333333333333,
"eval_table_size": 0,
"evals_per_epoch": 4,
"experimental_skip_move_to_device": true,
"expert_parallel_backend": "torch",
"expert_parallel_dispatch_chunks": 1,
"expert_parallel_fallback_on_unsupported": true,
"expert_parallel_num_nvl_bytes": 268435456,
"expert_parallel_num_rdma_bytes": 0,
"expert_parallel_save_dispatch": false,
"expert_parallel_size": 1,
"fa_varlen_kwargs_in_collator": true,
"fp16": false,
"fsdp_version": 2,
"generate_samples": false,
"generation_do_sample": true,
"generation_max_new_tokens": 50,
"generation_prompt_ratio": 0.5,
"generation_temperature": 0.7,
"gradient_accumulation_steps": 8,
"gradient_checkpointing": false,
"include_tkps": true,
"is_multimodal": true,
"label_balance_window_optim_steps": 1,
"layer_offloading": false,
"learning_rate": 0.0001,
"lisa_layers_attribute": "model.layers",
"load_best_model_at_end": false,
"load_in_4bit": false,
"load_in_8bit": false,
"local_rank": 0,
"logging_nan_inf_filter": true,
"logging_steps": 1,
"lora_alpha": 64,
"lora_dropout": 0.0,
"lora_embedding_kernel": false,
"lora_fp32_gradients": false,
"lora_mlp_kernel": false,
"lora_model_dir": "./outputs/glm53-flash-qlora-out/checkpoint-234",
"lora_o_kernel": false,
"lora_qkv_kernel": false,
"lora_r": 32,
"lora_target_modules": [
"q_a_proj",
"q_b_proj",
"kv_a_proj_with_mqa",
"kv_b_proj",
"q_proj",
"k_proj",
"v_proj",
"o_proj"
],
"loraplus_lr_embedding": 1e-06,
"lr_scheduler": "cosine",
"mean_resizing_embeddings": false,
"merge_compile": false,
"merge_compile_precision": "strict",
"merge_lora": true,
"merge_method": "memory_efficient",
"merge_num_threads": 2,
"micro_batch_size": 1,
"mixlora_jitter_noise": 0.0,
"mixlora_num_experts": 8,
"mixlora_router_aux_loss_coef": 0.01,
"mixlora_router_init_range": 0.02,
"mixlora_top_k": 2,
"model_config_type": "glm5_next",
"model_config_type_text": "glm5_next_text",
"nf4_backend": "bitsandbytes",
"num_epochs": 3.0,
"num_generation_samples": 3,
"optimizer": "adamw_torch_8bit",
"otel_metrics_host": "localhost",
"otel_metrics_port": 8000,
"output_dir": "./outputs/glm53-flash-qlora-out",
"ple_cpu_offload": false,
"plugins": [
"axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin"
],
"pretrain_multipack_attn": true,
"processor_config": "zai-org/GLM-5.3-Flash-BF16",
"profiler_steps_start": 0,
"qgalore_cos_threshold": 0.4,
"qgalore_gamma_proj": 2,
"qgalore_proj_bits": 4,
"qgalore_proj_group_size": 256,
"qgalore_proj_quant": true,
"qgalore_proj_type": "std",
"qgalore_queue_size": 5,
"qgalore_rank": 256,
"qgalore_scale": 0.25,
"qgalore_update_proj_gap": 200,
"qlora_sharded_model_loading": false,
"quantize_moe_experts": false,
"ray_num_workers": 1,
"relora_prune_method": "magnitude",
"resources_per_worker": {
"GPU": 1
},
"sample_packing": false,
"sample_packing_bin_size": 200,
"sample_packing_group_size": 100000,
"save_only_model": false,
"save_safetensors": true,
"save_steps": 0.3333333333333333,
"saves_per_epoch": 1,
"sequence_len": 3072,
"shuffle_before_merging_datasets": false,
"shuffle_merged_datasets": true,
"skip_prepare_dataset": false,
"streaming_multipack_buffer_size": 10000,
"strict": false,
"sync_profiler_start_step": 2,
"tensor_parallel_size": 1,
"tf32": false,
"tiled_mlp_use_original_mlp": true,
"tokenizer_config": "zai-org/GLM-5.3-Flash-BF16",
"tokenizer_save_jinja_files": true,
"torch_compile_regional": false,
"torch_dtype": "torch.bfloat16",
"train_on_inputs": false,
"trl": {
"async_prefetch": false,
"log_completions": false,
"mask_truncated_completions": false,
"ref_model_mixup_alpha": 0.9,
"ref_model_sync_steps": 64,
"replay_buffer_size": 0,
"replay_recompute_logps": true,
"reroll_max_groups": 1,
"reroll_start_fraction": 1.0,
"reward_num_workers": 1,
"scale_rewards": true,
"skip_zero_advantage_batches": true,
"sync_ref_model": false,
"use_data_producer": false,
"use_vllm": false,
"vllm_lora_sync": false,
"vllm_server_host": "0.0.0.0",
"vllm_server_port": 8000
},
"use_otel_metrics": false,
"use_ray": false,
"val_set_size": 0.1,
"vllm": {
"device": "auto",
"dtype": "auto",
"host": "0.0.0.0",
"port": 8000
},
"warmup_ratio": 0.1,
"weight_decay": 0.0,
"world_size": 1
}
[2026-11-05 11:42:40,442] [INFO] [axolotl.cli.merge_lora.do_merge_lora:36] [PID:985839] Using 2 CPU threads for LoRA merge
[2026-11-05 11:42:40,443] [DEBUG] [axolotl.cli.merge_lora._do_merge_lora:68] [PID:985839] Using memory-efficient LoRA merging method...
[2026-11-05 11:42:40,443] [DEBUG] [axolotl.cli.merge_lora._do_merge_lora_efficient:143] [PID:985839] Using memory-efficient LoRA merging method...
Downloading bytes: | 0.00B
Reconstructing (incomplete total...): | | 0.00B / 0.00B [A
Fetching 130 files: 0%| | 0/130 [00:00<?, ?it/s][A[A
Fetching 130 files: 90%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 117/130 [00:00<00:00, 1163.50it/s][A[A
Reconstructing (incomplete total...): 0%| | 0.00B / 8.03kB [A
Reconstructing (incomplete total...): 100%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 8.03kB / 8.03kB [ADownloading bytes: βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 8.03kB
Reconstructing (incomplete total...): 84%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 8.03kB / 9.60kB [A
Reconstructing (incomplete total...): 90%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 9.60kB / 10.7kB [AFetching 130 files: 100%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 130/130 [00:00<00:00, 490.11it/s]
Download complete: βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 10.7kB
Reconstruction complete: 100%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 10.7kB / 10.7kB [A[2026-11-05 11:42:41,626] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:2927] [PID:985839] Found 14 weight renamings for this model type
[2026-11-05 11:42:41,626] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:2930] [PID:985839] Found 3 weight converters (fuse/unfuse) for this model type. Will fuseβmergeβunfuse within each shard.
[2026-11-05 11:42:41,628] [DEBUG] [axolotl.cli.utils.lora_merge._build_layer_type_map:412] [PID:985839] Layer type map: 690 modules (36 conv layers)
[2026-11-05 11:42:41,628] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3021] [PID:985839] LoRA scale factor: 2.0 (rslora=False)
[2026-11-05 11:42:41,628] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3024] [PID:985839] NF4 simulation enabled: base weights will undergo quantizeβdequantize roundtrip before LoRA merge to match QLoRA training dynamics
[2026-11-05 11:42:41,628] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3037] [PID:985839] Loading LoRA weights from outputs/glm53-flash-qlora-out/checkpoint-234/adapter_model.safetensors
[2026-11-05 11:42:41,637] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3043] [PID:985839] Keeping LoRA weights on CPU; will move per-tensor during merge
[2026-11-05 11:42:41,641] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3091] [PID:985839] Found 120 model shards in /home/anon/.cache/huggingface/hub/models--zai-org--GLM-5.3-Flash-BF16/snapshots/f9fed3bbc82291073373d67452e9f25317ef7577
[2026-11-05 11:42:41,641] [INFO] [axolotl.cli.utils.lora_merge.copy_non_model_files:969] [PID:985839] Copying non-model files to output directory...
[2026-11-05 11:42:41,642] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying tokenizer.json to output
[2026-11-05 11:42:41,662] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying README.md to output
[2026-11-05 11:42:41,662] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying tokenizer_config.json to output
[2026-11-05 11:42:41,663] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying config.json to output
[2026-11-05 11:42:41,663] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying chat_template.jinja to output
[2026-11-05 11:42:41,664] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying .gitattributes to output
[2026-11-05 11:42:41,664] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying generation_config.json to output
[2026-11-05 11:42:41,665] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying LICENSE to output
[2026-11-05 11:42:41,665] [DEBUG] [axolotl.cli.utils.lora_merge.copy_non_model_files:997] [PID:985839] Copying processor_config.json to output
Merging shards: 0%| | 0/120 [00:00<?, ?it/s][A[A[2026-11-05 11:42:41,791] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.10.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [33, 33] experts per pattern, expected 288); carrying 66 per-expert tensors to the shard that completes the group
[2026-11-05 11:42:41,798] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.10.mlp.experts.down_proj': expert list incomplete in this shard (found [33] experts per pattern, expected 288); carrying 33 per-expert tensors to the shard that completes the group
[2026-11-05 11:42:42,475] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.0.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:42:43,026] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.0.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:42:43,376] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.0.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:42:43,752] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.0.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:42:44,441] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.1.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:42:44,798] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.1.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:42:45,151] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.1.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:42:45,511] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.1.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Download complete: βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 10.7kB
Reconstruction complete: 100%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 10.7kB / 10.7kB
Merging shards: 1%|β | 1/120 [00:04<09:01, 4.55s/it][A[A[2026-11-05 11:42:46,230] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.10.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [140, 139] experts per pattern, expected 288); carrying 279 per-expert tensors to the shard that completes the group
[2026-11-05 11:42:46,230] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.10.mlp.experts.down_proj': expert list incomplete in this shard (found [140] experts per pattern, expected 288); carrying 140 per-expert tensors to the shard that completes the group
[2026-11-05 11:42:46,230] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00002-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 2%|ββ | 2/120 [00:04<04:09, 2.12s/it][A[A[2026-11-05 11:42:46,718] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.10.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [246, 246] experts per pattern, expected 288); carrying 492 per-expert tensors to the shard that completes the group
[2026-11-05 11:42:46,720] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.10.mlp.experts.down_proj': expert list incomplete in this shard (found [247] experts per pattern, expected 288); carrying 247 per-expert tensors to the shard that completes the group
[2026-11-05 11:42:46,720] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00003-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 2%|βββ | 3/120 [00:05<02:39, 1.37s/it][A[A[2026-11-05 11:43:04,501] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.11.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [58, 58] experts per pattern, expected 288); carrying 116 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:11,388] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.11.mlp.experts.down_proj': expert list incomplete in this shard (found [59] experts per pattern, expected 288); carrying 59 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:11,461] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.10.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:43:11,814] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.10.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:43:12,175] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.10.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:43:12,541] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.10.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 3%|βββ | 4/120 [00:31<21:36, 11.17s/it][A[A[2026-11-05 11:43:13,341] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.11.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [165, 165] experts per pattern, expected 288); carrying 330 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:13,342] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.11.mlp.experts.down_proj': expert list incomplete in this shard (found [165] experts per pattern, expected 288); carrying 165 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:13,344] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00005-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 4%|ββββ | 5/120 [00:32<13:59, 7.30s/it][A[A[2026-11-05 11:43:13,813] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.11.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [272, 271] experts per pattern, expected 288); carrying 543 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:13,817] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.11.mlp.experts.down_proj': expert list incomplete in this shard (found [272] experts per pattern, expected 288); carrying 272 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:13,818] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00006-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 5%|βββββ | 6/120 [00:32<09:27, 4.98s/it][A[A[2026-11-05 11:43:32,781] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.12.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [84, 84] experts per pattern, expected 288); carrying 168 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:39,764] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.12.mlp.experts.down_proj': expert list incomplete in this shard (found [84] experts per pattern, expected 288); carrying 84 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:39,832] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.11.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:43:39,854] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.11.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:43:40,035] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.11.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:43:40,735] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.11.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:43:40,800] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.11.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 6%|ββββββ | 7/120 [00:59<23:06, 12.27s/it][A[A[2026-11-05 11:43:41,492] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.12.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [191, 190] experts per pattern, expected 288); carrying 381 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:41,495] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.12.mlp.experts.down_proj': expert list incomplete in this shard (found [191] experts per pattern, expected 288); carrying 191 per-expert tensors to the shard that completes the group
[2026-11-05 11:43:41,495] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00008-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 7%|βββββββ | 8/120 [01:00<15:50, 8.49s/it][A[A[2026-11-05 11:44:00,086] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.13.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [3, 2] experts per pattern, expected 288); carrying 5 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:06,918] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.13.mlp.experts.down_proj': expert list incomplete in this shard (found [3] experts per pattern, expected 288); carrying 3 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:06,987] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.12.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:44:07,362] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.12.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:44:07,746] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.12.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:44:08,108] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.12.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 8%|βββββββ | 9/120 [01:27<26:24, 14.28s/it][A[A[2026-11-05 11:44:08,998] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.13.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [109, 109] experts per pattern, expected 288); carrying 218 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:08,999] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.13.mlp.experts.down_proj': expert list incomplete in this shard (found [110] experts per pattern, expected 288); carrying 110 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:08,999] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00010-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 8%|ββββββββ | 10/120 [01:27<18:23, 10.03s/it][A[A[2026-11-05 11:44:09,406] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.13.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [216, 216] experts per pattern, expected 288); carrying 432 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:09,409] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.13.mlp.experts.down_proj': expert list incomplete in this shard (found [216] experts per pattern, expected 288); carrying 216 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:09,409] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00011-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 9%|βββββββββ | 11/120 [01:28<12:51, 7.08s/it][A[A[2026-11-05 11:44:28,767] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.14.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [28, 28] experts per pattern, expected 288); carrying 56 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:35,726] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.14.mlp.experts.down_proj': expert list incomplete in this shard (found [28] experts per pattern, expected 288); carrying 28 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:35,802] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.13.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:44:36,188] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.13.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:44:36,574] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.13.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:44:36,943] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.13.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 10%|ββββββββββ | 12/120 [01:56<24:10, 13.43s/it][A[A[2026-11-05 11:44:37,804] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.14.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [135, 134] experts per pattern, expected 288); carrying 269 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:37,805] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.14.mlp.experts.down_proj': expert list incomplete in this shard (found [135] experts per pattern, expected 288); carrying 135 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:37,805] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00013-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 11%|ββββββββββ | 13/120 [01:56<16:55, 9.49s/it][A[A[2026-11-05 11:44:38,252] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.14.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [241, 241] experts per pattern, expected 288); carrying 482 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:38,253] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.14.mlp.experts.down_proj': expert list incomplete in this shard (found [242] experts per pattern, expected 288); carrying 242 per-expert tensors to the shard that completes the group
[2026-11-05 11:44:38,253] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00014-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 12%|βββββββββββ | 14/120 [01:56<11:57, 6.76s/it][A[A[2026-11-05 11:44:57,285] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.15.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [53, 53] experts per pattern, expected 288); carrying 106 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:04,203] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.15.mlp.experts.down_proj': expert list incomplete in this shard (found [54] experts per pattern, expected 288); carrying 54 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:04,276] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.14.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:45:04,655] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.14.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:45:05,014] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.14.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:45:05,401] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.14.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 12%|ββββββββββββ | 15/120 [02:24<22:47, 13.02s/it][A[A[2026-11-05 11:45:06,182] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.15.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [160, 160] experts per pattern, expected 288); carrying 320 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:06,182] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.15.mlp.experts.down_proj': expert list incomplete in this shard (found [160] experts per pattern, expected 288); carrying 160 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:06,183] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00016-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 13%|βββββββββββββ | 16/120 [02:24<15:59, 9.23s/it][A[A[2026-11-05 11:45:06,678] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.15.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [267, 266] experts per pattern, expected 288); carrying 533 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:06,678] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.15.mlp.experts.down_proj': expert list incomplete in this shard (found [267] experts per pattern, expected 288); carrying 267 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:06,678] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00017-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 14%|βββββββββββββ | 17/120 [02:25<11:19, 6.60s/it][A[A[2026-11-05 11:45:25,646] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.16.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [79, 79] experts per pattern, expected 288); carrying 158 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:32,795] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.16.mlp.experts.down_proj': expert list incomplete in this shard (found [79] experts per pattern, expected 288); carrying 79 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:32,872] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.15.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:45:32,895] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.15.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:45:33,082] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.15.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:45:33,813] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.15.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:45:33,870] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.15.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 15%|ββββββββββββββ | 18/120 [02:52<21:55, 12.90s/it][A[A[2026-11-05 11:45:34,657] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.16.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [186, 185] experts per pattern, expected 288); carrying 371 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:34,658] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.16.mlp.experts.down_proj': expert list incomplete in this shard (found [186] experts per pattern, expected 288); carrying 186 per-expert tensors to the shard that completes the group
[2026-11-05 11:45:34,659] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00019-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 16%|βββββββββββββββ | 19/120 [02:53<15:26, 9.18s/it][A[A[2026-11-05 11:46:03,790] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.16.self_attn.conv1d.weight': expert list incomplete in this shard (found [1, 1] experts per pattern); carrying 2 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:03,857] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.16.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:46:04,247] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.16.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
Merging shards: 17%|ββββββββββββββββ | 20/120 [03:23<25:39, 15.40s/it][A[A[2026-11-05 11:46:05,143] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.17.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [104, 103] experts per pattern, expected 288); carrying 207 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:05,144] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.17.mlp.experts.down_proj': expert list incomplete in this shard (found [104] experts per pattern, expected 288); carrying 104 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:05,146] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.16.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:46:05,567] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.16.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 18%|ββββββββββββββββ | 21/120 [03:33<22:04, 13.38s/it][A[A[2026-11-05 11:46:15,570] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.17.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [210, 210] experts per pattern, expected 288); carrying 420 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:15,574] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.17.mlp.experts.down_proj': expert list incomplete in this shard (found [211] experts per pattern, expected 288); carrying 211 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:15,577] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00022-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 18%|βββββββββββββββββ | 22/120 [03:34<16:24, 10.05s/it][A[A[2026-11-05 11:46:34,359] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.18.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [22, 22] experts per pattern, expected 288); carrying 44 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:42,660] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.18.mlp.experts.down_proj': expert list incomplete in this shard (found [23] experts per pattern, expected 288); carrying 23 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:42,746] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.17.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:46:43,111] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.17.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:46:43,483] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.17.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:46:43,845] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.17.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 19%|ββββββββββββββββββ | 23/120 [04:03<25:17, 15.64s/it][A[A[2026-11-05 11:46:44,783] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.18.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [129, 129] experts per pattern, expected 288); carrying 258 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:44,784] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.18.mlp.experts.down_proj': expert list incomplete in this shard (found [129] experts per pattern, expected 288); carrying 129 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:44,785] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00024-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 20%|βββββββββββββββββββ | 24/120 [04:03<17:45, 11.10s/it][A[A[2026-11-05 11:46:45,171] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.18.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [236, 235] experts per pattern, expected 288); carrying 471 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:45,175] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.18.mlp.experts.down_proj': expert list incomplete in this shard (found [236] experts per pattern, expected 288); carrying 236 per-expert tensors to the shard that completes the group
[2026-11-05 11:46:45,175] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00025-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 21%|ββββββββββββββββββββ | 25/120 [04:03<12:29, 7.89s/it][A[A[2026-11-05 11:47:07,146] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.19.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [48, 47] experts per pattern, expected 288); carrying 95 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:14,981] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.19.mlp.experts.down_proj': expert list incomplete in this shard (found [48] experts per pattern, expected 288); carrying 48 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:15,049] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.18.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:47:15,423] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.18.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:47:15,782] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.18.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:47:16,151] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.18.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 22%|ββββββββββββββββββββ | 26/120 [04:35<23:24, 14.94s/it][A[A[2026-11-05 11:47:17,008] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.19.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [154, 154] experts per pattern, expected 288); carrying 308 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:17,037] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.19.mlp.experts.down_proj': expert list incomplete in this shard (found [155] experts per pattern, expected 288); carrying 155 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:17,061] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00027-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 22%|βββββββββββββββββββββ | 27/120 [04:35<16:26, 10.61s/it][A[A[2026-11-05 11:47:17,472] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.19.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [261, 261] experts per pattern, expected 288); carrying 522 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:17,475] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.19.mlp.experts.down_proj': expert list incomplete in this shard (found [261] experts per pattern, expected 288); carrying 261 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:17,475] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00028-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 23%|ββββββββββββββββββββββ | 28/120 [04:36<11:33, 7.54s/it][A[A[2026-11-05 11:47:36,837] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.20.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [62, 62] experts per pattern, expected 288); carrying 124 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:44,151] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.20.mlp.experts.down_proj': expert list incomplete in this shard (found [62] experts per pattern, expected 288); carrying 62 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:44,227] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.19.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:47:44,251] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.19.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:47:44,467] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.19.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:47:45,211] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.19.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:47:45,269] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.19.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
[2026-11-05 11:47:46,000] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.2.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:47:46,384] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.2.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:47:46,757] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.2.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:47:47,123] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.2.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 24%|βββββββββββββββββββββββ | 29/120 [05:06<21:40, 14.29s/it][A[A[2026-11-05 11:47:47,911] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.20.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [169, 168] experts per pattern, expected 288); carrying 337 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:47,912] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.20.mlp.experts.down_proj': expert list incomplete in this shard (found [169] experts per pattern, expected 288); carrying 169 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:47,913] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00030-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 25%|βββββββββββββββββββββββ | 30/120 [05:06<15:10, 10.12s/it][A[A[2026-11-05 11:47:48,559] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.20.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [275, 275] experts per pattern, expected 288); carrying 550 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:48,560] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.20.mlp.experts.down_proj': expert list incomplete in this shard (found [276] experts per pattern, expected 288); carrying 276 per-expert tensors to the shard that completes the group
[2026-11-05 11:47:48,560] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00031-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 26%|ββββββββββββββββββββββββ | 31/120 [05:07<10:48, 7.28s/it][A[A[2026-11-05 11:48:08,693] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.21.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [87, 87] experts per pattern, expected 288); carrying 174 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:15,883] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.21.mlp.experts.down_proj': expert list incomplete in this shard (found [88] experts per pattern, expected 288); carrying 88 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:15,957] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.20.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:48:16,341] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.20.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:48:16,723] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.20.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:48:17,094] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.20.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 27%|βββββββββββββββββββββββββ | 32/120 [05:36<20:12, 13.78s/it][A[A[2026-11-05 11:48:17,894] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.21.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [194, 194] experts per pattern, expected 288); carrying 388 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:17,894] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.21.mlp.experts.down_proj': expert list incomplete in this shard (found [194] experts per pattern, expected 288); carrying 194 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:17,895] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00033-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 28%|ββββββββββββββββββββββββββ | 33/120 [05:36<14:09, 9.76s/it][A[A[2026-11-05 11:48:37,293] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.22.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [6, 6] experts per pattern, expected 288); carrying 12 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:44,131] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.22.mlp.experts.down_proj': expert list incomplete in this shard (found [6] experts per pattern, expected 288); carrying 6 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:44,203] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.21.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:48:44,578] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.21.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:48:44,939] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.21.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:48:45,302] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.21.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 28%|ββββββββββββββββββββββββββ | 34/120 [06:04<21:45, 15.18s/it][A[A[2026-11-05 11:48:46,207] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.22.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [113, 112] experts per pattern, expected 288); carrying 225 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:46,208] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.22.mlp.experts.down_proj': expert list incomplete in this shard (found [113] experts per pattern, expected 288); carrying 113 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:46,208] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00035-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 29%|βββββββββββββββββββββββββββ | 35/120 [06:04<15:16, 10.78s/it][A[A[2026-11-05 11:48:46,619] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.22.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [219, 219] experts per pattern, expected 288); carrying 438 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:46,620] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.22.mlp.experts.down_proj': expert list incomplete in this shard (found [220] experts per pattern, expected 288); carrying 220 per-expert tensors to the shard that completes the group
[2026-11-05 11:48:46,620] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00036-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 30%|ββββββββββββββββββββββββββββ | 36/120 [06:05<10:43, 7.67s/it][A[A[2026-11-05 11:49:05,534] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.23.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [31, 31] experts per pattern, expected 288); carrying 62 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:12,780] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.23.mlp.experts.down_proj': expert list incomplete in this shard (found [32] experts per pattern, expected 288); carrying 32 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:12,850] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.22.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:49:13,203] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.22.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:49:13,582] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.22.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:49:13,933] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.22.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 31%|βββββββββββββββββββββββββββββ | 37/120 [06:33<18:55, 13.68s/it][A[A[2026-11-05 11:49:14,827] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.23.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [138, 138] experts per pattern, expected 288); carrying 276 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:14,828] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.23.mlp.experts.down_proj': expert list incomplete in this shard (found [138] experts per pattern, expected 288); carrying 138 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:14,828] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00038-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 32%|ββββββββββββββββββββββββββββββ | 38/120 [06:33<13:17, 9.72s/it][A[A[2026-11-05 11:49:15,215] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.23.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [245, 244] experts per pattern, expected 288); carrying 489 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:15,217] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.23.mlp.experts.down_proj': expert list incomplete in this shard (found [245] experts per pattern, expected 288); carrying 245 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:15,218] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00039-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 32%|ββββββββββββββββββββββββββββββ | 39/120 [06:33<09:21, 6.94s/it][A[A[2026-11-05 11:49:35,180] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.24.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [57, 57] experts per pattern, expected 288); carrying 114 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:42,578] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.24.mlp.experts.down_proj': expert list incomplete in this shard (found [57] experts per pattern, expected 288); carrying 57 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:42,653] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.23.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:49:42,680] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.23.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:49:42,861] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.23.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:49:43,557] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.23.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:49:43,613] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.23.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 33%|βββββββββββββββββββββββββββββββ | 40/120 [07:02<17:57, 13.46s/it][A[A[2026-11-05 11:49:44,373] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.24.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [164, 163] experts per pattern, expected 288); carrying 327 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:44,375] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.24.mlp.experts.down_proj': expert list incomplete in this shard (found [164] experts per pattern, expected 288); carrying 164 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:44,375] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00041-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 34%|ββββββββββββββββββββββββββββββββ | 41/120 [07:03<12:35, 9.56s/it][A[A[2026-11-05 11:49:44,886] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.24.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [270, 270] experts per pattern, expected 288); carrying 540 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:44,889] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.24.mlp.experts.down_proj': expert list incomplete in this shard (found [271] experts per pattern, expected 288); carrying 271 per-expert tensors to the shard that completes the group
[2026-11-05 11:49:44,889] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00042-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 35%|βββββββββββββββββββββββββββββββββ | 42/120 [07:03<08:53, 6.84s/it][A[A[2026-11-05 11:50:05,606] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.25.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [82, 82] experts per pattern, expected 288); carrying 164 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:12,571] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.25.mlp.experts.down_proj': expert list incomplete in this shard (found [83] experts per pattern, expected 288); carrying 83 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:12,639] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.24.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:50:13,002] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.24.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:50:13,372] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.24.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:50:13,782] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.24.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 36%|βββββββββββββββββββββββββββββββββ | 43/120 [07:32<17:25, 13.58s/it][A[A[2026-11-05 11:50:14,614] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.25.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [189, 189] experts per pattern, expected 288); carrying 378 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:14,615] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.25.mlp.experts.down_proj': expert list incomplete in this shard (found [189] experts per pattern, expected 288); carrying 189 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:14,615] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00044-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 37%|ββββββββββββββββββββββββββββββββββ | 44/120 [07:33<12:11, 9.63s/it][A[A[2026-11-05 11:50:33,588] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.26.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [1, 1] experts per pattern, expected 288); carrying 2 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:40,765] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.26.mlp.experts.down_proj': expert list incomplete in this shard (found [1] experts per pattern, expected 288); carrying 1 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:40,838] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.25.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:50:41,198] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.25.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:50:41,572] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.25.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:50:41,928] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.25.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 38%|βββββββββββββββββββββββββββββββββββ | 45/120 [08:01<18:49, 15.06s/it][A[A[2026-11-05 11:50:42,833] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.26.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [108, 107] experts per pattern, expected 288); carrying 215 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:42,834] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.26.mlp.experts.down_proj': expert list incomplete in this shard (found [108] experts per pattern, expected 288); carrying 108 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:42,834] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00046-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 38%|ββββββββββββββββββββββββββββββββββββ | 46/120 [08:01<13:10, 10.69s/it][A[A[2026-11-05 11:50:43,221] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.26.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [214, 214] experts per pattern, expected 288); carrying 428 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:43,226] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.26.mlp.experts.down_proj': expert list incomplete in this shard (found [215] experts per pattern, expected 288); carrying 215 per-expert tensors to the shard that completes the group
[2026-11-05 11:50:43,226] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00047-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 39%|ββββββββββββββββββββββββββββββββββββ | 47/120 [08:01<09:14, 7.60s/it][A[A[2026-11-05 11:51:01,227] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.27.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [26, 26] experts per pattern, expected 288); carrying 52 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:08,141] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.27.mlp.experts.down_proj': expert list incomplete in this shard (found [27] experts per pattern, expected 288); carrying 27 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:08,218] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.26.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:51:08,593] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.26.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:51:08,977] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.26.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:51:09,351] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.26.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 40%|βββββββββββββββββββββββββββββββββββββ | 48/120 [08:28<15:55, 13.28s/it][A[A[2026-11-05 11:51:10,188] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.27.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [133, 133] experts per pattern, expected 288); carrying 266 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:10,190] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.27.mlp.experts.down_proj': expert list incomplete in this shard (found [133] experts per pattern, expected 288); carrying 133 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:10,193] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00049-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 41%|ββββββββββββββββββββββββββββββββββββββ | 49/120 [08:28<11:09, 9.43s/it][A[A[2026-11-05 11:51:10,662] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.27.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [240, 239] experts per pattern, expected 288); carrying 479 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:10,664] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.27.mlp.experts.down_proj': expert list incomplete in this shard (found [240] experts per pattern, expected 288); carrying 240 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:10,665] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00050-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 42%|βββββββββββββββββββββββββββββββββββββββ | 50/120 [08:29<07:51, 6.74s/it][A[A[2026-11-05 11:51:29,113] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.28.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [52, 52] experts per pattern, expected 288); carrying 104 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:35,968] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.28.mlp.experts.down_proj': expert list incomplete in this shard (found [52] experts per pattern, expected 288); carrying 52 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:36,059] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.27.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:51:36,082] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.27.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:51:36,269] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.27.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:51:37,026] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.27.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:51:37,087] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.27.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 42%|βββββββββββββββββββββββββββββββββββββββ | 51/120 [08:56<14:39, 12.74s/it][A[A[2026-11-05 11:51:37,848] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.28.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [159, 158] experts per pattern, expected 288); carrying 317 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:37,849] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.28.mlp.experts.down_proj': expert list incomplete in this shard (found [159] experts per pattern, expected 288); carrying 159 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:37,850] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00052-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 43%|ββββββββββββββββββββββββββββββββββββββββ | 52/120 [08:56<10:15, 9.05s/it][A[A[2026-11-05 11:51:38,346] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.28.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [265, 265] experts per pattern, expected 288); carrying 530 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:38,350] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.28.mlp.experts.down_proj': expert list incomplete in this shard (found [266] experts per pattern, expected 288); carrying 266 per-expert tensors to the shard that completes the group
[2026-11-05 11:51:38,352] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00053-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 44%|βββββββββββββββββββββββββββββββββββββββββ | 53/120 [08:57<07:14, 6.49s/it][A[A[2026-11-05 11:51:56,710] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.29.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [77, 77] experts per pattern, expected 288); carrying 154 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:03,678] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.29.mlp.experts.down_proj': expert list incomplete in this shard (found [78] experts per pattern, expected 288); carrying 78 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:03,766] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.28.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:52:04,126] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.28.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:52:04,517] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.28.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:52:04,877] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.28.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 45%|ββββββββββββββββββββββββββββββββββββββββββ | 54/120 [09:24<13:53, 12.62s/it][A[A[2026-11-05 11:52:05,701] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.29.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [184, 184] experts per pattern, expected 288); carrying 368 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:05,703] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.29.mlp.experts.down_proj': expert list incomplete in this shard (found [184] experts per pattern, expected 288); carrying 184 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:05,703] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00055-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 46%|βββββββββββββββββββββββββββββββββββββββββββ | 55/120 [09:24<09:42, 8.96s/it][A[A[2026-11-05 11:52:32,041] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.29.self_attn.conv1d.weight': expert list incomplete in this shard (found [1] experts per pattern); carrying 1 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:32,116] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.29.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 47%|βββββββββββββββββββββββββββββββββββββββββββ | 56/120 [09:51<15:16, 14.32s/it][A[A[2026-11-05 11:52:33,030] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.3.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [102, 102] experts per pattern, expected 288); carrying 204 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:33,031] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.3.mlp.experts.down_proj': expert list incomplete in this shard (found [103] experts per pattern, expected 288); carrying 103 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:33,033] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.29.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:52:33,419] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.29.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:52:33,778] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.29.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 48%|ββββββββββββββββββββββββββββββββββββββββββββ | 57/120 [09:57<12:12, 11.63s/it][A[A[2026-11-05 11:52:40,016] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.3.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [209, 209] experts per pattern, expected 288); carrying 418 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:40,017] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.3.mlp.experts.down_proj': expert list incomplete in this shard (found [209] experts per pattern, expected 288); carrying 209 per-expert tensors to the shard that completes the group
[2026-11-05 11:52:40,020] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00058-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 48%|βββββββββββββββββββββββββββββββββββββββββββββ | 58/120 [09:58<09:05, 8.80s/it][A[A[2026-11-05 11:53:13,420] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.30.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [21, 21] experts per pattern, expected 288); carrying 42 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:20,348] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.30.mlp.experts.down_proj': expert list incomplete in this shard (found [22] experts per pattern, expected 288); carrying 22 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:20,436] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.3.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:53:20,459] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.3.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:53:20,636] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.3.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:53:21,337] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.3.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:53:21,410] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.3.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 49%|ββββββββββββββββββββββββββββββββββββββββββββββ | 59/120 [10:40<18:57, 18.64s/it][A[A[2026-11-05 11:53:22,131] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.30.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [128, 128] experts per pattern, expected 288); carrying 256 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:22,133] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.30.mlp.experts.down_proj': expert list incomplete in this shard (found [128] experts per pattern, expected 288); carrying 128 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:22,134] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00060-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 50%|ββββββββββββββββββββββββββββββββββββββββββββββ | 60/120 [10:40<13:11, 13.19s/it][A[A[2026-11-05 11:53:22,640] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.30.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [235, 234] experts per pattern, expected 288); carrying 469 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:22,642] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.30.mlp.experts.down_proj': expert list incomplete in this shard (found [235] experts per pattern, expected 288); carrying 235 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:22,642] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00061-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 51%|βββββββββββββββββββββββββββββββββββββββββββββββ | 61/120 [10:41<09:13, 9.38s/it][A[A[2026-11-05 11:53:48,657] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.31.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [47, 46] experts per pattern, expected 288); carrying 93 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:56,616] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.31.mlp.experts.down_proj': expert list incomplete in this shard (found [47] experts per pattern, expected 288); carrying 47 per-expert tensors to the shard that completes the group
[2026-11-05 11:53:56,705] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.30.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:53:57,097] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.30.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:53:57,480] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.30.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:53:57,844] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.30.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 52%|ββββββββββββββββββββββββββββββββββββββββββββββββ | 62/120 [11:19<17:20, 17.93s/it][A[A[2026-11-05 11:54:00,930] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.31.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [153, 153] experts per pattern, expected 288); carrying 306 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:00,930] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.31.mlp.experts.down_proj': expert list incomplete in this shard (found [154] experts per pattern, expected 288); carrying 154 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:00,930] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00063-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 52%|βββββββββββββββββββββββββββββββββββββββββββββββββ | 63/120 [11:19<12:02, 12.68s/it][A[A[2026-11-05 11:54:01,462] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.31.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [260, 260] experts per pattern, expected 288); carrying 520 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:01,464] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.31.mlp.experts.down_proj': expert list incomplete in this shard (found [260] experts per pattern, expected 288); carrying 260 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:01,465] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00064-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 53%|βββββββββββββββββββββββββββββββββββββββββββββββββ | 64/120 [11:20<08:25, 9.03s/it][A[A[2026-11-05 11:54:27,433] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.32.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [72, 72] experts per pattern, expected 288); carrying 144 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:34,416] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.32.mlp.experts.down_proj': expert list incomplete in this shard (found [73] experts per pattern, expected 288); carrying 73 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:34,494] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.31.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:54:34,522] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.31.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:54:34,703] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.31.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:54:35,420] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.31.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:54:35,478] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.31.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 54%|ββββββββββββββββββββββββββββββββββββββββββββββββββ | 65/120 [11:54<15:13, 16.61s/it][A[A[2026-11-05 11:54:36,369] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.32.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [179, 179] experts per pattern, expected 288); carrying 358 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:36,381] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.32.mlp.experts.down_proj': expert list incomplete in this shard (found [179] experts per pattern, expected 288); carrying 179 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:36,383] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00066-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 55%|βββββββββββββββββββββββββββββββββββββββββββββββββββ | 66/120 [11:55<10:37, 11.81s/it][A[A[2026-11-05 11:54:37,773] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.32.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [286, 285] experts per pattern, expected 288); carrying 571 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:37,780] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.32.mlp.experts.down_proj': expert list incomplete in this shard (found [286] experts per pattern, expected 288); carrying 286 per-expert tensors to the shard that completes the group
[2026-11-05 11:54:37,780] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00067-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 56%|ββββββββββββββββββββββββββββββββββββββββββββββββββββ | 67/120 [11:56<07:40, 8.69s/it][A[A[2026-11-05 11:55:04,647] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.33.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [98, 97] experts per pattern, expected 288); carrying 195 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:11,676] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.33.mlp.experts.down_proj': expert list incomplete in this shard (found [98] experts per pattern, expected 288); carrying 98 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:11,745] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.32.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:55:12,100] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.32.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:55:12,475] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.32.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:55:12,855] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.32.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 57%|βββββββββββββββββββββββββββββββββββββββββββββββββββββ | 68/120 [12:32<14:30, 16.73s/it][A[A[2026-11-05 11:55:13,677] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.33.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [204, 204] experts per pattern, expected 288); carrying 408 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:13,677] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.33.mlp.experts.down_proj': expert list incomplete in this shard (found [205] experts per pattern, expected 288); carrying 205 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:13,677] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00069-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 57%|βββββββββββββββββββββββββββββββββββββββββββββββββββββ | 69/120 [12:32<10:03, 11.83s/it][A[A[2026-11-05 11:55:32,953] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.34.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [16, 16] experts per pattern, expected 288); carrying 32 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:40,078] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.34.mlp.experts.down_proj': expert list incomplete in this shard (found [17] experts per pattern, expected 288); carrying 17 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:40,164] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.33.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:55:40,557] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.33.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:55:40,933] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.33.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:55:41,299] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.33.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 58%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 70/120 [13:00<13:55, 16.72s/it][A[A[2026-11-05 11:55:42,463] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.34.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [123, 123] experts per pattern, expected 288); carrying 246 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:42,467] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.34.mlp.experts.down_proj': expert list incomplete in this shard (found [123] experts per pattern, expected 288); carrying 123 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:42,467] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00071-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 59%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 71/120 [13:01<09:43, 11.91s/it][A[A[2026-11-05 11:55:42,921] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.34.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [230, 229] experts per pattern, expected 288); carrying 459 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:42,923] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.34.mlp.experts.down_proj': expert list incomplete in this shard (found [230] experts per pattern, expected 288); carrying 230 per-expert tensors to the shard that completes the group
[2026-11-05 11:55:42,923] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00072-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 60%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 72/120 [13:01<06:46, 8.47s/it][A[A[2026-11-05 11:56:05,860] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.35.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [42, 41] experts per pattern, expected 288); carrying 83 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:13,455] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.35.mlp.experts.down_proj': expert list incomplete in this shard (found [42] experts per pattern, expected 288); carrying 42 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:13,609] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.34.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:56:13,977] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.34.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:56:14,349] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.34.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:56:14,774] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.34.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 61%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 73/120 [13:33<12:13, 15.60s/it][A[A[2026-11-05 11:56:15,577] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.35.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [148, 148] experts per pattern, expected 288); carrying 296 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:15,578] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.35.mlp.experts.down_proj': expert list incomplete in this shard (found [149] experts per pattern, expected 288); carrying 149 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:15,578] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00074-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 62%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 74/120 [13:34<08:27, 11.04s/it][A[A[2026-11-05 11:56:16,081] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.35.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [255, 255] experts per pattern, expected 288); carrying 510 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:16,083] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.35.mlp.experts.down_proj': expert list incomplete in this shard (found [255] experts per pattern, expected 288); carrying 255 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:16,083] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00075-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 62%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 75/120 [13:34<05:54, 7.89s/it][A[A[2026-11-05 11:56:42,719] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.36.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [67, 67] experts per pattern, expected 288); carrying 134 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:59,529] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.36.mlp.experts.down_proj': expert list incomplete in this shard (found [68] experts per pattern, expected 288); carrying 68 per-expert tensors to the shard that completes the group
[2026-11-05 11:56:59,632] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.35.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:56:59,659] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.35.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:56:59,852] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.35.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:57:00,630] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.35.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:57:00,697] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.35.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 63%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 76/120 [14:19<13:55, 19.00s/it][A[A[2026-11-05 11:57:01,421] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.36.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [174, 174] experts per pattern, expected 288); carrying 348 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:01,422] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.36.mlp.experts.down_proj': expert list incomplete in this shard (found [174] experts per pattern, expected 288); carrying 174 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:01,422] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00077-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 64%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 77/120 [14:20<09:37, 13.42s/it][A[A[2026-11-05 11:57:01,942] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.36.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [281, 280] experts per pattern, expected 288); carrying 561 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:01,946] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.36.mlp.experts.down_proj': expert list incomplete in this shard (found [281] experts per pattern, expected 288); carrying 281 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:01,947] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00078-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 65%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 78/120 [14:20<06:41, 9.56s/it][A[A[2026-11-05 11:57:19,964] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.37.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [93, 92] experts per pattern, expected 288); carrying 185 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:27,032] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.37.mlp.experts.down_proj': expert list incomplete in this shard (found [93] experts per pattern, expected 288); carrying 93 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:27,113] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.36.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:57:27,482] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.36.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:57:27,851] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.36.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:57:28,229] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.36.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 66%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 79/120 [14:47<10:02, 14.69s/it][A[A[2026-11-05 11:57:29,028] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.37.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [199, 199] experts per pattern, expected 288); carrying 398 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:29,029] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.37.mlp.experts.down_proj': expert list incomplete in this shard (found [200] experts per pattern, expected 288); carrying 200 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:29,029] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00080-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 67%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 80/120 [14:47<06:56, 10.40s/it][A[A[2026-11-05 11:57:50,556] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.38.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [11, 11] experts per pattern, expected 288); carrying 22 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:57,433] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.38.mlp.experts.down_proj': expert list incomplete in this shard (found [12] experts per pattern, expected 288); carrying 12 per-expert tensors to the shard that completes the group
[2026-11-05 11:57:57,785] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.37.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:57:58,515] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.37.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:57:59,130] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.37.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:57:59,764] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.37.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 68%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 81/120 [15:19<10:51, 16.69s/it][A[A[2026-11-05 11:58:00,944] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.38.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [118, 118] experts per pattern, expected 288); carrying 236 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:00,948] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.38.mlp.experts.down_proj': expert list incomplete in this shard (found [118] experts per pattern, expected 288); carrying 118 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:00,948] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00082-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 68%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 82/120 [15:19<07:30, 11.85s/it][A[A[2026-11-05 11:58:02,148] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.38.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [225, 224] experts per pattern, expected 288); carrying 449 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:02,158] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.38.mlp.experts.down_proj': expert list incomplete in this shard (found [225] experts per pattern, expected 288); carrying 225 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:02,159] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00083-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 69%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 83/120 [15:20<05:20, 8.67s/it][A[A[2026-11-05 11:58:29,640] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.39.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [37, 36] experts per pattern, expected 288); carrying 73 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:36,469] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.39.mlp.experts.down_proj': expert list incomplete in this shard (found [37] experts per pattern, expected 288); carrying 37 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:36,713] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.38.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:58:37,329] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.38.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:58:37,936] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.38.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:58:38,527] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.38.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 70%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 84/120 [15:57<10:17, 17.17s/it][A[A[2026-11-05 11:58:39,641] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.39.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [143, 143] experts per pattern, expected 288); carrying 286 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:39,645] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.39.mlp.experts.down_proj': expert list incomplete in this shard (found [144] experts per pattern, expected 288); carrying 144 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:39,647] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00085-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 71%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 85/120 [15:58<07:05, 12.15s/it][A[A[2026-11-05 11:58:40,892] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.39.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [250, 250] experts per pattern, expected 288); carrying 500 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:40,901] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.39.mlp.experts.down_proj': expert list incomplete in this shard (found [250] experts per pattern, expected 288); carrying 250 per-expert tensors to the shard that completes the group
[2026-11-05 11:58:40,901] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00086-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 72%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 86/120 [15:59<05:01, 8.88s/it][A[A[2026-11-05 11:59:00,442] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.4.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [62, 62] experts per pattern, expected 288); carrying 124 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:07,301] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.4.mlp.experts.down_proj': expert list incomplete in this shard (found [63] experts per pattern, expected 288); carrying 63 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:07,405] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.39.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 11:59:07,428] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.39.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 11:59:07,616] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.39.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 11:59:08,381] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.39.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 11:59:08,439] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.39.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 72%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 87/120 [16:27<08:00, 14.57s/it][A[A[2026-11-05 11:59:09,146] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.4.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [169, 169] experts per pattern, expected 288); carrying 338 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:09,147] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.4.mlp.experts.down_proj': expert list incomplete in this shard (found [169] experts per pattern, expected 288); carrying 169 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:09,147] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00088-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 73%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 88/120 [16:27<05:30, 10.32s/it][A[A[2026-11-05 11:59:09,645] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.4.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [276, 275] experts per pattern, expected 288); carrying 551 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:09,647] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.4.mlp.experts.down_proj': expert list incomplete in this shard (found [276] experts per pattern, expected 288); carrying 276 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:09,648] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00089-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 74%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 89/120 [16:28<03:48, 7.37s/it][A[A[2026-11-05 11:59:30,796] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.40.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [88, 87] experts per pattern, expected 288); carrying 175 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:37,659] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.40.mlp.experts.down_proj': expert list incomplete in this shard (found [88] experts per pattern, expected 288); carrying 88 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:37,753] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.4.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:59:38,128] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.4.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 11:59:38,517] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.4.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 11:59:38,897] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.4.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 75%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 90/120 [16:58<07:01, 14.06s/it][A[A[2026-11-05 11:59:39,689] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.40.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [194, 194] experts per pattern, expected 288); carrying 388 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:39,690] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.40.mlp.experts.down_proj': expert list incomplete in this shard (found [195] experts per pattern, expected 288); carrying 195 per-expert tensors to the shard that completes the group
[2026-11-05 11:59:39,690] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00091-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 76%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 91/120 [16:58<04:48, 9.96s/it][A[A[2026-11-05 11:59:59,462] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.41.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [6, 6] experts per pattern, expected 288); carrying 12 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:06,164] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.41.mlp.experts.down_proj': expert list incomplete in this shard (found [7] experts per pattern, expected 288); carrying 7 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:06,235] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.40.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:00:06,621] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.40.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:00:06,976] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.40.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:00:07,334] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.40.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 77%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 92/120 [17:26<07:10, 15.39s/it][A[A[2026-11-05 12:00:08,253] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.41.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [113, 113] experts per pattern, expected 288); carrying 226 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:08,253] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.41.mlp.experts.down_proj': expert list incomplete in this shard (found [113] experts per pattern, expected 288); carrying 113 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:08,253] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00093-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 78%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 93/120 [17:27<04:55, 10.94s/it][A[A[2026-11-05 12:00:08,689] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.41.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [220, 219] experts per pattern, expected 288); carrying 439 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:08,689] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.41.mlp.experts.down_proj': expert list incomplete in this shard (found [220] experts per pattern, expected 288); carrying 220 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:08,689] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00094-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 78%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 94/120 [17:27<03:22, 7.78s/it][A[A[2026-11-05 12:00:38,942] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.42.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [32, 31] experts per pattern, expected 288); carrying 63 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:46,049] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.42.mlp.experts.down_proj': expert list incomplete in this shard (found [32] experts per pattern, expected 288); carrying 32 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:46,124] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.41.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:00:46,476] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.41.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:00:46,831] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.41.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:00:47,184] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.41.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 79%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 95/120 [18:06<07:07, 17.11s/it][A[A[2026-11-05 12:00:47,982] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.42.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [138, 138] experts per pattern, expected 288); carrying 276 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:47,983] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.42.mlp.experts.down_proj': expert list incomplete in this shard (found [139] experts per pattern, expected 288); carrying 139 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:47,983] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00096-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 80%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 96/120 [18:06<04:50, 12.11s/it][A[A[2026-11-05 12:00:48,544] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.42.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [245, 245] experts per pattern, expected 288); carrying 490 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:48,546] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.42.mlp.experts.down_proj': expert list incomplete in this shard (found [245] experts per pattern, expected 288); carrying 245 per-expert tensors to the shard that completes the group
[2026-11-05 12:00:48,546] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00097-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 81%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 97/120 [18:07<03:18, 8.64s/it][A[A[2026-11-05 12:01:10,294] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.43.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [57, 57] experts per pattern, expected 288); carrying 114 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:17,238] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.43.mlp.experts.down_proj': expert list incomplete in this shard (found [57] experts per pattern, expected 288); carrying 57 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:17,302] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.42.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:01:17,708] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.42.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:01:18,068] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.42.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:01:18,429] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.42.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 82%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 98/120 [18:37<05:32, 15.13s/it][A[A[2026-11-05 12:01:19,240] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.43.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [164, 163] experts per pattern, expected 288); carrying 327 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:19,242] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.43.mlp.experts.down_proj': expert list incomplete in this shard (found [164] experts per pattern, expected 288); carrying 164 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:19,242] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00099-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 82%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 99/120 [18:37<03:45, 10.72s/it][A[A[2026-11-05 12:01:20,312] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.43.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [270, 270] experts per pattern, expected 288); carrying 540 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:20,367] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.43.mlp.experts.down_proj': expert list incomplete in this shard (found [271] experts per pattern, expected 288); carrying 271 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:20,373] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00100-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 83%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 100/120 [18:39<02:37, 7.85s/it][A[A[2026-11-05 12:01:44,529] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.44.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [83, 82] experts per pattern, expected 288); carrying 165 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:51,363] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.44.mlp.experts.down_proj': expert list incomplete in this shard (found [83] experts per pattern, expected 288); carrying 83 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:51,482] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.43.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
[2026-11-05 12:01:51,505] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.43.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 12:01:51,691] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.43.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 12:01:52,396] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.43.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 12:01:52,504] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.43.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 84%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 101/120 [19:11<04:48, 15.21s/it][A[A[2026-11-05 12:01:53,184] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.44.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [189, 189] experts per pattern, expected 288); carrying 378 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:53,185] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.44.mlp.experts.down_proj': expert list incomplete in this shard (found [190] experts per pattern, expected 288); carrying 190 per-expert tensors to the shard that completes the group
[2026-11-05 12:01:53,185] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00102-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 85%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 102/120 [19:11<03:13, 10.77s/it][A[A[2026-11-05 12:02:26,319] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.44.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:02:26,933] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.44.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:02:27,652] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.44.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:02:28,344] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.44.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 86%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 103/120 [19:47<05:11, 18.30s/it][A[A[2026-11-05 12:02:29,822] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.45.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [107, 106] experts per pattern, expected 288); carrying 213 per-expert tensors to the shard that completes the group
[2026-11-05 12:02:29,847] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.45.mlp.experts.down_proj': expert list incomplete in this shard (found [107] experts per pattern, expected 288); carrying 107 per-expert tensors to the shard that completes the group
[2026-11-05 12:02:29,857] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00104-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 87%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 104/120 [19:48<03:28, 13.06s/it][A[A[2026-11-05 12:02:30,726] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.45.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [213, 213] experts per pattern, expected 288); carrying 426 per-expert tensors to the shard that completes the group
[2026-11-05 12:02:30,780] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.45.mlp.experts.down_proj': expert list incomplete in this shard (found [214] experts per pattern, expected 288); carrying 214 per-expert tensors to the shard that completes the group
[2026-11-05 12:02:30,807] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00105-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 88%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 105/120 [19:49<02:21, 9.42s/it][A[A[2026-11-05 12:02:56,149] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.5.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [26, 25] experts per pattern, expected 288); carrying 51 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:03,048] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.5.mlp.experts.down_proj': expert list incomplete in this shard (found [26] experts per pattern, expected 288); carrying 26 per-expert tensors to the shard that completes the group
Merging shards: 88%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 106/120 [20:21<03:47, 16.28s/it][A[A[2026-11-05 12:03:03,509] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.5.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [132, 132] experts per pattern, expected 288); carrying 264 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:03,509] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.5.mlp.experts.down_proj': expert list incomplete in this shard (found [133] experts per pattern, expected 288); carrying 133 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:03,510] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00107-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 89%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 107/120 [20:22<02:29, 11.51s/it][A[A[2026-11-05 12:03:04,008] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.5.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [239, 239] experts per pattern, expected 288); carrying 478 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:04,010] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.5.mlp.experts.down_proj': expert list incomplete in this shard (found [239] experts per pattern, expected 288); carrying 239 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:04,010] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00108-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 90%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 108/120 [20:22<01:38, 8.21s/it][A[A[2026-11-05 12:03:25,031] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.6.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [51, 51] experts per pattern, expected 288); carrying 102 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:32,108] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.6.mlp.experts.down_proj': expert list incomplete in this shard (found [51] experts per pattern, expected 288); carrying 51 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:32,173] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.5.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:03:32,557] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.5.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:03:32,924] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.5.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:03:33,298] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.5.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 91%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 109/120 [20:52<02:41, 14.65s/it][A[A[2026-11-05 12:03:34,131] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.6.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [158, 157] experts per pattern, expected 288); carrying 315 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:34,132] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.6.mlp.experts.down_proj': expert list incomplete in this shard (found [158] experts per pattern, expected 288); carrying 158 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:34,160] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00110-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 92%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 110/120 [20:52<01:44, 10.40s/it][A[A[2026-11-05 12:03:35,065] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.6.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [264, 264] experts per pattern, expected 288); carrying 528 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:35,070] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.6.mlp.experts.down_proj': expert list incomplete in this shard (found [265] experts per pattern, expected 288); carrying 265 per-expert tensors to the shard that completes the group
[2026-11-05 12:03:35,071] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00111-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 92%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 111/120 [20:53<01:08, 7.56s/it][A[A[2026-11-05 12:04:04,303] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.7.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [76, 76] experts per pattern, expected 288); carrying 152 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:11,297] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.7.mlp.experts.down_proj': expert list incomplete in this shard (found [77] experts per pattern, expected 288); carrying 77 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:11,374] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.6.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:04:11,744] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.6.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:04:12,110] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.6.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:04:12,504] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.6.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 93%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 112/120 [21:31<02:13, 16.64s/it][A[A[2026-11-05 12:04:13,364] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.7.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [183, 183] experts per pattern, expected 288); carrying 366 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:13,368] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.7.mlp.experts.down_proj': expert list incomplete in this shard (found [183] experts per pattern, expected 288); carrying 183 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:13,368] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00113-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 94%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 113/120 [21:32<01:22, 11.79s/it][A[A[2026-11-05 12:04:39,185] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.7.self_attn.kv_a_proj_with_mqa.weight: torch.Size([32, 4096]), torch.Size([512, 32])
Merging shards: 95%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 114/120 [21:57<01:36, 16.00s/it][A[A[2026-11-05 12:04:39,745] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.8.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [102, 102] experts per pattern, expected 288); carrying 204 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:39,745] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.8.mlp.experts.down_proj': expert list incomplete in this shard (found [102] experts per pattern, expected 288); carrying 102 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:39,746] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.7.self_attn.kv_b_proj.weight: torch.Size([32, 512]), torch.Size([32768, 32])
[2026-11-05 12:04:39,944] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.7.self_attn.o_proj.weight: torch.Size([32, 16384]), torch.Size([4096, 32])
[2026-11-05 12:04:40,699] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.7.self_attn.q_a_proj.weight: torch.Size([32, 4096]), torch.Size([1536, 32])
[2026-11-05 12:04:40,758] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.7.self_attn.q_b_proj.weight: torch.Size([32, 1536]), torch.Size([16384, 32])
Merging shards: 96%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 115/120 [22:04<01:05, 13.14s/it][A[A[2026-11-05 12:04:46,658] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.8.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [209, 208] experts per pattern, expected 288); carrying 417 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:46,661] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.8.mlp.experts.down_proj': expert list incomplete in this shard (found [209] experts per pattern, expected 288); carrying 209 per-expert tensors to the shard that completes the group
[2026-11-05 12:04:46,662] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00116-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 97%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 116/120 [22:05<00:37, 9.49s/it][A[A[2026-11-05 12:05:11,354] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.9.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [21, 20] experts per pattern, expected 288); carrying 41 per-expert tensors to the shard that completes the group
[2026-11-05 12:05:18,324] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.9.mlp.experts.down_proj': expert list incomplete in this shard (found [21] experts per pattern, expected 288); carrying 21 per-expert tensors to the shard that completes the group
[2026-11-05 12:05:18,402] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.8.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:05:18,783] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.8.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:05:19,157] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.8.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:05:19,689] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.8.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 98%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 117/120 [22:39<00:50, 16.74s/it][A[A[2026-11-05 12:05:20,725] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.9.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [127, 127] experts per pattern, expected 288); carrying 254 per-expert tensors to the shard that completes the group
[2026-11-05 12:05:20,726] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.9.mlp.experts.down_proj': expert list incomplete in this shard (found [128] experts per pattern, expected 288); carrying 128 per-expert tensors to the shard that completes the group
[2026-11-05 12:05:20,726] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00118-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 98%|ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | 118/120 [22:39<00:23, 11.84s/it][A[A[2026-11-05 12:05:21,295] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.9.mlp.experts.gate_up_proj': expert list incomplete in this shard (found [234, 234] experts per pattern, expected 288); carrying 468 per-expert tensors to the shard that completes the group
[2026-11-05 12:05:21,297] [INFO] [axolotl.cli.utils.lora_merge._fuse_and_unfuse_with_merge:2717] [PID:985839] Deferring fuse for 'model.language_model.layers.9.mlp.experts.down_proj': expert list incomplete in this shard (found [234] experts per pattern, expected 288); carrying 234 per-expert tensors to the shard that completes the group
[2026-11-05 12:05:21,297] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3363] [PID:985839] model-00119-of-00120.safetensors holds only expert tensors carried to a later shard; writing no file
Merging shards: 99%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 119/120 [22:40<00:08, 8.46s/it][A[A[2026-11-05 12:05:47,367] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.9.self_attn.k_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:05:47,757] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.9.self_attn.o_proj.weight: torch.Size([32, 8192]), torch.Size([4096, 32])
[2026-11-05 12:05:48,127] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.9.self_attn.q_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
[2026-11-05 12:05:48,507] [DEBUG] [axolotl.cli.utils.lora_merge._merge_tensor_with_lora:2300] [PID:985839] Merging LoRA for model.language_model.layers.9.self_attn.v_proj.weight: torch.Size([32, 4096]), torch.Size([8192, 32])
Merging shards: 100%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 120/120 [23:09<00:00, 14.71s/it][A[AMerging shards: 100%|βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ| 120/120 [23:09<00:00, 11.58s/it]
[2026-11-05 12:05:58,571] [DEBUG] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3430] [PID:985839] Wrote weight-map index: model.safetensors.index.json
[2026-11-05 12:05:58,571] [INFO] [axolotl.cli.utils.lora_merge.merge_lora_sharded_efficient:3452] [PID:985839] Applied LoRA to 191/38770 tensors
[2026-11-05 12:05:58,571] [INFO] [axolotl.utils.lora_merge_precision.log_merge_retention:181] [PID:985839] LoRA merge rounding: merge-lora: 99.8% of the LoRA delta survives (73.5% of weight entries change) across 191 merged LoRA linears
[2026-11-05 12:05:58,576] [DEBUG] [axolotl.cli.merge_lora._do_merge_lora_efficient:214] [PID:985839] Memory-efficient LoRA merge completed successfully!
[2026-11-05 11:42:39,984] [DEBUG] [axolotl.utils.config.resolve_dtype:168] [PID:985839] bf16 support detected, enabling for this configuration.
[2026-11-05 11:42:40,199] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:985839] baseline 0.000GB ()
[2026-11-05 11:42:40,200] [INFO] [axolotl.cli.config.load_cfg:339] [PID:985839] config:
{
"activation_offloading": true,
"adapter": "qlora",
"adapter_checkpoint_format": "gather",
"attn_decontaminates_packing": true,
"attn_implementation": "sdpa",
"attn_needs_dtype_cast": false,
"attn_supports_packing": false,
"attn_uses_flash_lib": false,
"axolotl_config_path": "glm53flash-merge.yml",
"balance_labels": false,
"base_model": "zai-org/GLM-5.3-Flash-BF16",
"base_model_config": "zai-org/GLM-5.3-Flash-BF16",
"batch_size": 8,
"bf16": true,
"capabilities": {
"bf16": true,
"compute_capability": "sm_120",
"fp8": true,
"n_gpu": 1,
"n_node": 1,
"tf32": true
},
"context_parallel_size": 1,
"cut_cross_entropy": true,
"cut_cross_entropy_accum_c_fp32": false,
"dataloader_num_workers": 1,
"dataloader_pin_memory": true,
"dataloader_prefetch_factor": 256,
"dataset_num_proc": 64,
"dataset_prepared_path": "last_run_prepared",
"datasets": [
{
"chat_template": "tokenizer_default",
"message_property_mappings": {
"content": "content",
"role": "role"
},
"path": "/home/anon/LoRA/Moni_messages.json",
"trust_remote_code": false,
"type": "chat_template"
}
],
"ddp": false,
"device": "cuda:0",
"dion_rank_fraction": 1.0,
"dion_rank_multiple_of": 1,
"eaft_alpha": 1.0,
"eaft_k": 20,
"env_capabilities": {
"torch_version": "2.13.0"
},
"eval_batch_size": 1,
"eval_causal_lm_metrics": [
"sacrebleu",
"comet",
"ter",
"chrf"
],
"eval_max_new_tokens": 128,
"eval_sample_packing": false,
"eval_steps": 0.08333333333333333,
"eval_table_size": 0,
"evals_per_epoch": 4,
"experimental_skip_move_to_device": true,
"expert_parallel_backend": "torch",
"expert_parallel_dispatch_chunks": 1,
"expert_parallel_fallback_on_unsupported": true,
"expert_parallel_num_nvl_bytes": 268435456,
"expert_parallel_num_rdma_bytes": 0,
"expert_parallel_save_dispatch": false,
"expert_parallel_size": 1,
"fa_varlen_kwargs_in_collator": true,
"fp16": false,
"fsdp_config": {
"activation_checkpointing": true,
"auto_wrap_policy": "TRANSFORMER_BASED_WRAP",
"cpu_ram_efficient_loading": true,
"fsdp_version": 2,
"offload_params": false,
"reshard_after_forward": true,
"state_dict_type": "FULL_STATE_DICT",
"transformer_layer_cls_to_wrap": "Glm5NextTextDecoderLayer"
},
"fsdp_version": 2,
"generate_samples": false,
"generation_do_sample": true,
"generation_max_new_tokens": 50,
"generation_prompt_ratio": 0.5,
"generation_temperature": 0.7,
"gradient_accumulation_steps": 8,
"gradient_checkpointing": false,
"include_tkps": true,
"is_multimodal": true,
"label_balance_window_optim_steps": 1,
"layer_offloading": false,
"learning_rate": 0.0001,