hamishivi/swerl_qwen35_27b_fp32lm_dppo_tmax_step240
2.65M • Updated • 162
Hamish Ivison
hamishivi
AI & ML interests
NLP :)
Organizations
hamishivi/swerl_qwen35_27b_fp32lm_dppo_tmax_step200
2.65M • Updated • 102
hamishivi/swerl_qwen35_27b_fp32lm_dppo_tmax_step160
2.65M • Updated • 198
hamishivi/swerl_qwen35_4b_fp32lm_dppo
Updated • 464
hamishivi/swerl_qwen35_2b_fp32lm_dppo
Updated • 254
hamishivi/swerl_qwen35_9b_fp32lm_dppo_swesmith
Updated • 216
hamishivi/qwen35_9b_tmax_skill_tax_no_tool_call_sft
9B • Updated • 84
hamishivi/swerl_qwen35_27b_fp32lm_dppo_tmax_step100
2.65M • Updated • 125
hamishivi/Qwen3.5-2B
Image-Text-to-Text • 2B • Updated • 2.89k
hamishivi/swerl_qwen35_9b_fp32lm_dppo_cli
Updated • 520
hamishivi/swerl_qwen35_9b_fp32lm_dppo_endless
Updated • 580
hamishivi/swerl_qwen35_9b_fp32lm_dppo_termigen
Updated • 588
hamishivi/swerl_qwen35_9b_fp32lm_dppo_terminaltraj
Updated • 673
hamishivi/qwen3_5_9b_sft_ablations_bc_only_v1_sanitized
9B • Updated • 115
hamishivi/qwen3_5_9b_sft_scientific_minimax
9B • Updated • 76
hamishivi/swerl_qwen35_9b_fp32lm_dppo_openthoughts
Updated • 755
hamishivi/swerl_qwen35_9b_fp32lm_dppo_g32_step500
9B • Updated • 173
hamishivi/swerl_qwen35_9b_fp32lm_dppo_g32_step200
9B • Updated • 780
hamishivi/swerl_qwen35_9b_fp32lm_dppo_g32_step300
9B • Updated • 135
hamishivi/swerl_qwen35_9b_fp32lm_dppo_g32_step100
9B • Updated • 138
hamishivi/swerl_qwen35_9b_fp32lm_dppo_g32_step400
9B • Updated • 138
hamishivi/tmax-qwen35-9b-tmax-sft-full-20260513-65k
9B • Updated • 98
hamishivi/swerl_qwen35_9b_fp32lm_datamix_step100
9B • Updated • 119
hamishivi/swerl_qwen35_9b_fp32lm_b001_testrun_step140
9B • Updated • 116
hamishivi/swerl_qwen35_9b_fp32lm_b001_testrun_step100
9B • Updated • 119 • 1
hamishivi/swerl_qwen35_9b_fp32lm_b001_testrun_step160
9B • Updated • 1.36k
hamishivi/swerl_qwen35_9b_fp32lm_datamix_step200
9B • Updated • 154
hamishivi/swerl_qwen35_9b_fp32lm_datamix_step300
9B • Updated • 123
hamishivi/qwen3_5_9b_sft_ablations_redsearcher_sft
9B • Updated • 1.21k
hamishivi/tmax-qwen36-27b-tmax-sft-full-20260513-epoch1
27B • Updated • 9