x86-to-llvm-o0_epoch1-AMD

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

x86-to-llvm-o0

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-1.5B-Instruct on the x86-to-llvm-o0_part_00, the x86-to-llvm-o0_part_01, the x86-to-llvm-o0_part_02, the x86-to-llvm-o0_part_03, the x86-to-llvm-o0_part_04, the x86-to-llvm-o0_part_05, the x86-to-llvm-o0_part_06, the x86-to-llvm-o0_part_07, the x86-to-llvm-o0_part_08, the x86-to-llvm-o0_part_09, the x86-to-llvm-o0_part_10, the x86-to-llvm-o0_part_11, the x86-to-llvm-o0_part_12, the x86-to-llvm-o0_part_13, the x86-to-llvm-o0_part_14, the x86-to-llvm-o0_part_15, the x86-to-llvm-o0_part_16, the x86-to-llvm-o0_part_17, the x86-to-llvm-o0_part_18, the x86-to-llvm-o0_part_19, the x86-to-llvm-o0_part_20, the x86-to-llvm-o0_part_21, the x86-to-llvm-o0_part_22, the x86-to-llvm-o0_part_23, the x86-to-llvm-o0_part_24, the x86-to-llvm-o0_part_25, the x86-to-llvm-o0_part_26, the x86-to-llvm-o0_part_27, the x86-to-llvm-o0_part_28, the x86-to-llvm-o0_part_29, the x86-to-llvm-o0_part_30, the x86-to-llvm-o0_part_31, the x86-to-llvm-o0_part_32, the x86-to-llvm-o0_part_33, the x86-to-llvm-o0_part_34, the x86-to-llvm-o0_part_35, the x86-to-llvm-o0_part_36, the x86-to-llvm-o0_part_37, the x86-to-llvm-o0_part_38, the x86-to-llvm-o0_part_39, the x86-to-llvm-o0_part_40, the x86-to-llvm-o0_part_41, the x86-to-llvm-o0_part_42, the x86-to-llvm-o0_part_43, the x86-to-llvm-o0_part_44, the x86-to-llvm-o0_part_45 and the x86-to-llvm-o0_part_46 datasets.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• total_eval_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1.0

Training results

Framework versions

• Transformers 4.52.1
• Pytorch 2.8.0+rocm6.3
• Datasets 3.6.0
• Tokenizers 0.21.1