Upload folder using huggingface_hub

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-full-traj-grpo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+600

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-full-traj-grpo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "7gv0wygq",
+  "name": "qwen-2.5-1.5b-q1-full-traj-grpo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-full-traj-rloo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-full-traj-rloo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "pe3eoziw",
+  "name": "qwen-2.5-1.5b-q1-full-traj-rloo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-token-level-ppo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-token-level-ppo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "xon1runt",
+  "name": "qwen-2.5-1.5b-q1-token-level-ppo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-turn-level-ppo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q1-turn-level-ppo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "28jbmpl3",
+  "name": "qwen-2.5-1.5b-q1-turn-level-ppo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-full-traj-grpo/profiling/profiling_metrics.jsonl

@@ -0,0 +1,6 @@
+{"global_step": 1, "profiling/generation_peak_memory_allocated_gb": 58.27460241317749, "profiling/generation_peak_memory_reserved_gb": 58.724609375, "profiling/generation_cuda_synced_time_s": 4.503549575805664, "global_seqlen/min": 116984, "global_seqlen/max": 125980, "global_seqlen/minmax_diff": 8996, "global_seqlen/balanced_min": 121122, "global_seqlen/balanced_max": 121123, "global_seqlen/mean": 121122.5, "actor/entropy": 0.37764230370521545, "profiling/adv_cuda_synced_time_s": 0.035688161849975586, "actor/kl_loss": 0.0, "actor/kl_coef": 0.001, "actor/pg_loss": -0.0018075834959745407, "actor/pg_clipfrac": 0.0, "actor/ppo_kl": 0.0, "actor/pg_clipfrac_lower": 0.0, "actor/grad_norm": 0.0027486986946314573, "perf/mfu/actor": 0.277339519608334, "perf/actor_update_cuda_synced_time_s": 4.930273532867432, "perf/max_memory_allocated_gb": 77.46703100204468, "perf/max_memory_reserved_gb": 80.8828125, "perf/cpu_memory_used_gb": 146.80777740478516, "actor/lr": 1e-06, "training/global_step": 1, "training/epoch": 0, "critic/score/mean": 0.00390625, "critic/score/max": 1.0, "critic/score/min": 0.0, "critic/rewards/mean": 0.00390625, "critic/rewards/max": 1.0, "critic/rewards/min": 0.0, "critic/advantages/mean": -3.734873174621178e-11, "critic/advantages/max": 1.4999970197677612, "critic/advantages/min": -1.4999970197677612, "critic/returns/mean": -3.734873174621178e-11, "critic/returns/max": 1.4999970197677612, "critic/returns/min": -1.4999970197677612, "response_length/mean": 818.26953125, "response_length/max": 1555.0, "response_length/min": 584.0, "response_length/clip_ratio": 0.0, "response_length_non_aborted/mean": 818.26953125, "response_length_non_aborted/max": 1555.0, "response_length_non_aborted/min": 584.0, "response_length_non_aborted/clip_ratio": 0.0, "response/aborted_ratio": 0.0, "prompt_length/mean": 128.0, "prompt_length/max": 128.0, "prompt_length/min": 128.0, "prompt_length/clip_ratio": 0.0, "timing_s/start_profile": 0.00023628491908311844, "timing_s/generate_sequences": 4.44061803817749, "timing_s/generation_timing/max": 4.552060604095459, "timing_s/generation_timing/min": 4.288355350494385, "timing_s/generation_timing/topk_ratio": 0.125, "timing_s/gen": 7.005054711829871, "timing_s/reward": 0.20261325314641, "timing_s/old_log_prob": 2.115626527927816, "timing_s/ref": 2.829029892105609, "timing_s/adv": 0.03592432010918856, "timing_s/update_actor": 5.2934510936029255, "timing_s/step": 17.514541666023433, "timing_s/stop_profile": 4.1150953620672226e-05, "timing_per_token_ms/gen": 0.008360171655873761, "timing_per_token_ms/adv": 3.707436697268112e-05, "timing_per_token_ms/update_actor": 0.005462910579787947, "timing_per_token_ms/ref": 0.0029195957523432984, "perf/total_num_tokens": 968980, "perf/time_per_step": 17.514541666023433, "perf/throughput": 6915.539230750542, "profiling/advantage/numel": 2097152, "profiling/advantage/shape": "[1024, 2048]", "profiling/advantage/batch_size": 1024, "profiling/advantage/seq_len": 2048, "profiling/throughput/cumulative_episodes": 1024, "profiling/throughput/elapsed_wall_clock_s": 21.648696899414062, "profiling/throughput/episodes_per_second": 47.30076848310049}
+{"global_step": 2, "profiling/generation_peak_memory_allocated_gb": 59.74354934692383, "profiling/generation_peak_memory_reserved_gb": 60.419921875, "profiling/generation_cuda_synced_time_s": 4.618285179138184, "global_seqlen/min": 115180, "global_seqlen/max": 126432, "global_seqlen/minmax_diff": 11252, "global_seqlen/balanced_min": 119555, "global_seqlen/balanced_max": 119556, "global_seqlen/mean": 119555.5, "actor/entropy": 0.39160576462745667, "profiling/adv_cuda_synced_time_s": 0.024133682250976562, "actor/kl_loss": 0.00037511294067371637, "actor/kl_coef": 0.001, "actor/pg_loss": 0.0019167001591995358, "actor/pg_clipfrac": 0.0, "actor/ppo_kl": 0.0, "actor/pg_clipfrac_lower": 0.0, "actor/grad_norm": 0.004408125299960375, "perf/mfu/actor": 0.2812501414178292, "perf/actor_update_cuda_synced_time_s": 4.796753883361816, "perf/max_memory_allocated_gb": 78.69405889511108, "perf/max_memory_reserved_gb": 81.95703125, "perf/cpu_memory_used_gb": 146.85517120361328, "actor/lr": 1e-06, "training/global_step": 2, "training/epoch": 0, "critic/score/mean": 0.005859375, "critic/score/max": 1.0, "critic/score/min": 0.0, "critic/rewards/mean": 0.005859375, "critic/rewards/max": 1.0, "critic/rewards/min": 0.0, "critic/advantages/mean": 2.3746471455865503e-10, "critic/advantages/max": 1.4999970197677612, "critic/advantages/min": -0.8660238981246948, "critic/returns/mean": 2.3746471455865503e-10, "critic/returns/max": 1.4999970197677612, "critic/returns/min": -0.8660238981246948, "response_length/mean": 806.02734375, "response_length/max": 1233.0, "response_length/min": 560.0, "response_length/clip_ratio": 0.0, "response_length_non_aborted/mean": 806.02734375, "response_length_non_aborted/max": 1233.0, "response_length_non_aborted/min": 560.0, "response_length_non_aborted/clip_ratio": 0.0, "response/aborted_ratio": 0.0, "prompt_length/mean": 128.0, "prompt_length/max": 128.0, "prompt_length/min": 128.0, "prompt_length/clip_ratio": 0.0, "timing_s/start_profile": 6.741425022482872e-05, "timing_s/generate_sequences": 4.47065544128418, "timing_s/generation_timing/max": 4.714422702789307, "timing_s/generation_timing/min": 4.352760314941406, "timing_s/generation_timing/topk_ratio": 0.125, "timing_s/gen": 7.863798919133842, "timing_s/reward": 0.19707466941326857, "timing_s/old_log_prob": 1.6557577135972679, "timing_s/ref": 1.443223049864173, "timing_s/adv": 0.02432147227227688, "timing_s/update_actor": 5.059991952031851, "timing_s/step": 16.273798420093954, "timing_s/stop_profile": 3.7575140595436096e-05, "timing_per_token_ms/gen": 0.009527581404668249, "timing_per_token_ms/adv": 2.5429060428291545e-05, "timing_per_token_ms/update_actor": 0.005290421553203168, "timing_per_token_ms/ref": 0.001508946733801637, "perf/total_num_tokens": 956444, "perf/time_per_step": 16.273798420093954, "perf/throughput": 7346.502452210525, "profiling/advantage/numel": 2097152, "profiling/advantage/shape": "[1024, 2048]", "profiling/advantage/batch_size": 1024, "profiling/advantage/seq_len": 2048, "profiling/throughput/cumulative_episodes": 2048, "profiling/throughput/elapsed_wall_clock_s": 37.96043682098389, "profiling/throughput/episodes_per_second": 53.95090708934888}
+{"global_step": 3, "profiling/generation_peak_memory_allocated_gb": 59.74354934692383, "profiling/generation_peak_memory_reserved_gb": 60.42578125, "profiling/generation_cuda_synced_time_s": 4.511974573135376, "global_seqlen/min": 119816, "global_seqlen/max": 125272, "global_seqlen/minmax_diff": 5456, "global_seqlen/balanced_min": 122462, "global_seqlen/balanced_max": 122463, "global_seqlen/mean": 122462.5, "actor/entropy": 0.3752768039703369, "profiling/adv_cuda_synced_time_s": 0.03005814552307129, "actor/kl_loss": 0.0004484600904106628, "actor/kl_coef": 0.001, "actor/pg_loss": -0.0008383354870602489, "actor/pg_clipfrac": 0.0, "actor/ppo_kl": 0.0, "actor/pg_clipfrac_lower": 0.0, "actor/grad_norm": 0.004709955770522356, "perf/mfu/actor": 0.2814664490102266, "perf/actor_update_cuda_synced_time_s": 4.9143359661102295, "perf/max_memory_allocated_gb": 79.13839435577393, "perf/max_memory_reserved_gb": 82.40625, "perf/cpu_memory_used_gb": 147.01629638671875, "actor/lr": 1e-06, "training/global_step": 3, "training/epoch": 0, "critic/score/mean": 0.01171875, "critic/score/max": 1.0, "critic/score/min": 0.0, "critic/rewards/mean": 0.01171875, "critic/rewards/max": 1.0, "critic/rewards/min": 0.0, "critic/advantages/mean": 3.067344711560338e-10, "critic/advantages/max": 1.4999970197677612, "critic/advantages/min": -1.4999970197677612, "critic/returns/mean": 3.067344711560338e-10, "critic/returns/max": 1.4999970197677612, "critic/returns/min": -1.4999970197677612, "response_length/mean": 828.73828125, "response_length/max": 1452.0, "response_length/min": 576.0, "response_length/clip_ratio": 0.0, "response_length_non_aborted/mean": 828.73828125, "response_length_non_aborted/max": 1452.0, "response_length_non_aborted/min": 576.0, "response_length_non_aborted/clip_ratio": 0.0, "response/aborted_ratio": 0.0, "prompt_length/mean": 128.0, "prompt_length/max": 128.0, "prompt_length/min": 128.0, "prompt_length/clip_ratio": 0.0, "timing_s/start_profile": 6.453227251768112e-05, "timing_s/generate_sequences": 4.475427150726318, "timing_s/generation_timing/max": 4.786633491516113, "timing_s/generation_timing/min": 4.316335201263428, "timing_s/generation_timing/topk_ratio": 0.125, "timing_s/gen": 8.025513769127429, "timing_s/reward": 0.2008807398378849, "timing_s/old_log_prob": 1.6254086419939995, "timing_s/ref": 1.4063659640960395, "timing_s/adv": 0.03022186690941453, "timing_s/update_actor": 5.188497003167868, "timing_s/step": 16.50952672632411, "timing_s/stop_profile": 3.419769927859306e-05, "timing_per_token_ms/gen": 0.009457045689191763, "timing_per_token_ms/adv": 3.084808299419672e-05, "timing_per_token_ms/update_actor": 0.005296005923413154, "timing_per_token_ms/ref": 0.0014355067511442681, "perf/total_num_tokens": 979700, "perf/time_per_step": 16.50952672632411, "perf/throughput": 7417.6868925464705, "profiling/advantage/numel": 2097152, "profiling/advantage/shape": "[1024, 2048]", "profiling/advantage/batch_size": 1024, "profiling/advantage/seq_len": 2048, "profiling/throughput/cumulative_episodes": 3072, "profiling/throughput/elapsed_wall_clock_s": 54.50162220001221, "profiling/throughput/episodes_per_second": 56.365294756296485}
+{"global_step": 4, "profiling/generation_peak_memory_allocated_gb": 59.74354934692383, "profiling/generation_peak_memory_reserved_gb": 60.419921875, "profiling/generation_cuda_synced_time_s": 4.527415752410889, "global_seqlen/min": 116660, "global_seqlen/max": 125504, "global_seqlen/minmax_diff": 8844, "global_seqlen/balanced_min": 121059, "global_seqlen/balanced_max": 121060, "global_seqlen/mean": 121059.5, "actor/entropy": 0.38037803769111633, "profiling/adv_cuda_synced_time_s": 0.03042459487915039, "actor/kl_loss": 0.0003841696452582255, "actor/kl_coef": 0.001, "actor/pg_loss": 0.00017056090291589499, "actor/pg_clipfrac": 0.0, "actor/ppo_kl": 0.0, "actor/pg_clipfrac_lower": 0.0, "actor/grad_norm": 0.004323871340602636, "perf/mfu/actor": 0.2816800293969281, "perf/actor_update_cuda_synced_time_s": 4.852255821228027, "perf/max_memory_allocated_gb": 78.91696834564209, "perf/max_memory_reserved_gb": 82.232421875, "perf/cpu_memory_used_gb": 148.08916473388672, "actor/lr": 1e-06, "training/global_step": 4, "training/epoch": 1, "critic/score/mean": 0.0048828125, "critic/score/max": 1.0, "critic/score/min": 0.0, "critic/rewards/mean": 0.0048828125, "critic/rewards/max": 1.0, "critic/rewards/min": 0.0, "critic/advantages/mean": 3.74100445066361e-10, "critic/advantages/max": 1.4999970197677612, "critic/advantages/min": -0.4999989867210388, "critic/returns/mean": 3.74100445066361e-10, "critic/returns/max": 1.4999970197677612, "critic/returns/min": -0.4999989867210388, "response_length/mean": 817.77734375, "response_length/max": 1509.0, "response_length/min": 560.0, "response_length/clip_ratio": 0.0, "response_length_non_aborted/mean": 817.77734375, "response_length_non_aborted/max": 1509.0, "response_length_non_aborted/min": 560.0, "response_length_non_aborted/clip_ratio": 0.0, "response/aborted_ratio": 0.0, "prompt_length/mean": 128.0, "prompt_length/max": 128.0, "prompt_length/min": 128.0, "prompt_length/clip_ratio": 0.0, "timing_s/start_profile": 0.00021633924916386604, "timing_s/generate_sequences": 4.410902500152588, "timing_s/generation_timing/max": 4.541614532470703, "timing_s/generation_timing/min": 4.292865753173828, "timing_s/generation_timing/topk_ratio": 0.125, "timing_s/gen": 8.127055485267192, "timing_s/reward": 0.20277301827445626, "timing_s/old_log_prob": 1.675968783441931, "timing_s/ref": 1.4407873549498618, "timing_s/adv": 0.030622732359915972, "timing_s/update_actor": 5.121905113104731, "timing_s/step": 16.65120204957202, "timing_s/stop_profile": 4.0865037590265274e-05, "timing_per_token_ms/gen": 0.009705059308609933, "timing_per_token_ms/adv": 3.1619505656222735e-05, "timing_per_token_ms/update_actor": 0.0052886236861881255, "timing_per_token_ms/ref": 0.0014876851413456419, "perf/total_num_tokens": 968476, "perf/time_per_step": 16.65120204957202, "perf/throughput": 7270.315959147919, "profiling/advantage/numel": 2097152, "profiling/advantage/shape": "[1024, 2048]", "profiling/advantage/batch_size": 1024, "profiling/advantage/seq_len": 2048, "profiling/throughput/cumulative_episodes": 4096, "profiling/throughput/elapsed_wall_clock_s": 71.73534107208252, "profiling/throughput/episodes_per_second": 57.09877361402906}
+{"global_step": 5, "profiling/generation_peak_memory_allocated_gb": 59.74354934692383, "profiling/generation_peak_memory_reserved_gb": 60.427734375, "profiling/generation_cuda_synced_time_s": 4.442465543746948, "global_seqlen/min": 116476, "global_seqlen/max": 127656, "global_seqlen/minmax_diff": 11180, "global_seqlen/balanced_min": 121348, "global_seqlen/balanced_max": 121349, "global_seqlen/mean": 121348.5, "actor/entropy": 0.4007837772369385, "profiling/adv_cuda_synced_time_s": 0.030620098114013672, "actor/kl_loss": 0.00041386188968317583, "actor/kl_coef": 0.001, "actor/pg_loss": -0.0003088494995608926, "actor/pg_clipfrac": 0.0, "actor/ppo_kl": 0.0, "actor/pg_clipfrac_lower": 0.0, "actor/grad_norm": 0.0026759228203445673, "perf/mfu/actor": 0.28181679937028903, "perf/actor_update_cuda_synced_time_s": 4.861763954162598, "perf/max_memory_allocated_gb": 78.96580123901367, "perf/max_memory_reserved_gb": 82.23046875, "perf/cpu_memory_used_gb": 148.09423828125, "actor/lr": 1e-06, "val-core/textworld_w2-o3-q4_1-1000/reward/mean@1": 0.01, "val-aux/success_rate": 0.01, "training/global_step": 5, "training/epoch": 1, "critic/score/mean": 0.00390625, "critic/score/max": 1.0, "critic/score/min": 0.0, "critic/rewards/mean": 0.00390625, "critic/rewards/max": 1.0, "critic/rewards/min": 0.0, "critic/advantages/mean": 9.963999764162068e-11, "critic/advantages/max": 1.4999970197677612, "critic/advantages/min": -0.8660238981246948, "critic/returns/mean": 9.963999764162068e-11, "critic/returns/max": 1.4999970197677612, "critic/returns/min": -0.8660238981246948, "response_length/mean": 820.03515625, "response_length/max": 1387.0, "response_length/min": 572.0, "response_length/clip_ratio": 0.0, "response_length_non_aborted/mean": 820.03515625, "response_length_non_aborted/max": 1387.0, "response_length_non_aborted/min": 572.0, "response_length_non_aborted/clip_ratio": 0.0, "response/aborted_ratio": 0.0, "prompt_length/mean": 128.0, "prompt_length/max": 128.0, "prompt_length/min": 128.0, "prompt_length/clip_ratio": 0.0, "timing_s/start_profile": 6.635533645749092e-05, "timing_s/generate_sequences": 4.454852104187012, "timing_s/generation_timing/max": 4.545742988586426, "timing_s/generation_timing/min": 4.29553747177124, "timing_s/generation_timing/topk_ratio": 0.125, "timing_s/gen": 7.740286306943744, "timing_s/reward": 0.19776281714439392, "timing_s/old_log_prob": 1.6279893810860813, "timing_s/ref": 1.4070004615932703, "timing_s/adv": 0.030794290360063314, "timing_s/update_actor": 5.116364544257522, "timing_s/step": 16.155977562069893, "timing_s/testing": 4.479943762067705, "timing_s/stop_profile": 3.938516601920128e-05, "timing_per_token_ms/gen": 0.009217743030910146, "timing_per_token_ms/adv": 3.1720921931527086e-05, "timing_per_token_ms/update_actor": 0.00527032116616349, "timing_per_token_ms/ref": 0.001449338538994374, "perf/total_num_tokens": 970788, "perf/time_per_step": 16.155977562069893, "perf/throughput": 7511.05895844367, "profiling/advantage/numel": 2097152, "profiling/advantage/shape": "[1024, 2048]", "profiling/advantage/batch_size": 1024, "profiling/advantage/seq_len": 2048, "profiling/throughput/cumulative_episodes": 5120, "profiling/throughput/elapsed_wall_clock_s": 92.42460608482361, "profiling/throughput/episodes_per_second": 55.396503343504314}
+{"global_step": 6, "profiling/generation_peak_memory_allocated_gb": 59.74354934692383, "profiling/generation_peak_memory_reserved_gb": 60.435546875, "profiling/generation_cuda_synced_time_s": 4.48847770690918, "global_seqlen/min": 113264, "global_seqlen/max": 127416, "global_seqlen/minmax_diff": 14152, "global_seqlen/balanced_min": 121339, "global_seqlen/balanced_max": 121340, "global_seqlen/mean": 121339.5, "actor/entropy": 0.39071711897850037, "profiling/adv_cuda_synced_time_s": 0.035332441329956055, "actor/kl_loss": 0.00037443024484673515, "actor/kl_coef": 0.001, "actor/pg_loss": 0.001176500867586583, "actor/pg_clipfrac": 0.0, "actor/ppo_kl": 0.0, "actor/pg_clipfrac_lower": 0.0, "actor/grad_norm": 0.0032005144748836756, "perf/mfu/actor": 0.2824457417996979, "perf/actor_update_cuda_synced_time_s": 4.850998163223267, "perf/max_memory_allocated_gb": 78.9670238494873, "perf/max_memory_reserved_gb": 82.232421875, "perf/cpu_memory_used_gb": 148.0805435180664, "actor/lr": 1e-06, "training/global_step": 6, "training/epoch": 1, "critic/score/mean": 0.0078125, "critic/score/max": 1.0, "critic/score/min": 0.0, "critic/rewards/mean": 0.0078125, "critic/rewards/max": 1.0, "critic/rewards/min": 0.0, "critic/advantages/mean": 2.2017328238366218e-10, "critic/advantages/max": 1.4999970197677612, "critic/advantages/min": -0.4999989867210388, "critic/returns/mean": 2.2017328238366218e-10, "critic/returns/max": 1.4999970197677612, "critic/returns/min": -0.4999989867210388, "response_length/mean": 819.96484375, "response_length/max": 1690.0, "response_length/min": 576.0, "response_length/clip_ratio": 0.0, "response_length_non_aborted/mean": 819.96484375, "response_length_non_aborted/max": 1690.0, "response_length_non_aborted/min": 576.0, "response_length_non_aborted/clip_ratio": 0.0, "response/aborted_ratio": 0.0, "prompt_length/mean": 128.0, "prompt_length/max": 128.0, "prompt_length/min": 128.0, "prompt_length/clip_ratio": 0.0, "timing_s/start_profile": 7.692910730838776e-05, "timing_s/generate_sequences": 4.433015823364258, "timing_s/generation_timing/max": 4.563940525054932, "timing_s/generation_timing/min": 4.28267765045166, "timing_s/generation_timing/topk_ratio": 0.125, "timing_s/gen": 6.916489327326417, "timing_s/reward": 0.20239367382600904, "timing_s/old_log_prob": 1.6656939540989697, "timing_s/ref": 1.4656466711312532, "timing_s/adv": 0.035506575368344784, "timing_s/update_actor": 5.092373747844249, "timing_s/step": 15.420811416115612, "timing_s/stop_profile": 3.6363955587148666e-05, "timing_per_token_ms/gen": 0.008237406957384816, "timing_per_token_ms/adv": 3.657771724000097e-05, "timing_per_token_ms/update_actor": 0.0052459975397997445, "timing_per_token_ms/ref": 0.0015098614539486865, "perf/total_num_tokens": 970716, "perf/time_per_step": 15.420811416115612, "perf/throughput": 7868.55482022129, "profiling/advantage/numel": 2097152, "profiling/advantage/shape": "[1024, 2048]", "profiling/advantage/batch_size": 1024, "profiling/advantage/seq_len": 2048, "profiling/throughput/cumulative_episodes": 6144, "profiling/throughput/elapsed_wall_clock_s": 107.8878664970398, "profiling/throughput/episodes_per_second": 56.94801648680835}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-full-traj-rloo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-full-traj-rloo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "i3ijveba",
+  "name": "qwen-2.5-1.5b-q2-full-traj-rloo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-token-level-ppo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-token-level-ppo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "v5p1hylj",
+  "name": "qwen-2.5-1.5b-q2-token-level-ppo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-turn-level-ppo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q2-turn-level-ppo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "1xhwzobu",
+  "name": "qwen-2.5-1.5b-q2-turn-level-ppo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q4-full-traj-rloo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q4-full-traj-rloo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "8yctgj6k",
+  "name": "qwen-2.5-1.5b-q4-full-traj-rloo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q4-token-level-ppo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+300

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q4-token-level-ppo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "gv6imluk",
+  "name": "qwen-2.5-1.5b-q4-token-level-ppo"
+}

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q4-token-level-rloo/latest_checkpointed_iteration.txt

@@ -0,0 +1 @@
+600

checkpoints/multi-turn-credit-assign/qwen-2.5-1.5b-q4-token-level-rloo/wandb_run_info.json

@@ -0,0 +1,4 @@
+{
+  "id": "06b4lkbw",
+  "name": "qwen-2.5-1.5b-q4-token-level-rloo"
+}