HridaAI/Hrida-T2SQL-3B-V0.2

1{2    "architectures": [3        "Phi3ForCausalLM"4    ],5    "attention_bias": false,6    "attention_dropout": 0.0,7    "auto_map": {8        "AutoConfig": "configuration_phi3.Phi3Config",9        "AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM"10    },11    "bos_token_id": 1,12    "embd_pdrop": 0.0,13    "eos_token_id": 32000,14    "hidden_act": "silu",15    "hidden_size": 3072,16    "initializer_range": 0.02,17    "intermediate_size": 8192,18    "max_position_embeddings": 131072,19    "model_type": "phi3",20    "num_attention_heads": 32,21    "num_hidden_layers": 32,22    "num_key_value_heads": 32,23    "original_max_position_embeddings": 4096,24    "pad_token_id": 32000,25    "resid_pdrop": 0.0,26    "rms_norm_eps": 1e-05,27    "rope_scaling": {28        "long_factor": [29            1.0800000429153442,30            1.1100000143051147,31            1.1399999856948853,32            1.340000033378601,33            1.5899999141693115,34            1.600000023841858,35            1.6200000047683716,36            2.620000123977661,37            3.2300000190734863,38            3.2300000190734863,39            4.789999961853027,40            7.400000095367432,41            7.700000286102295,42            9.09000015258789,43            12.199999809265137,44            17.670000076293945,45            24.46000099182129,46            28.57000160217285,47            30.420001983642578,48            30.840002059936523,49            32.590003967285156,50            32.93000411987305,51            42.320003509521484,52            44.96000289916992,53            50.340003967285156,54            50.45000457763672,55            57.55000305175781,56            57.93000411987305,57            58.21000289916992,58            60.1400032043457,59            62.61000442504883,60            62.62000274658203,61            62.71000289916992,62            63.1400032043457,63            63.1400032043457,64            63.77000427246094,65            63.93000411987305,66            63.96000289916992,67            63.970001220703125,68            64.02999877929688,69            64.06999969482422,70            64.08000183105469,71            64.12000274658203,72            64.41000366210938,73            64.4800033569336,74            64.51000213623047,75            64.52999877929688,76            64.8399963378906277        ],78        "short_factor": [79            1.0,80            1.0199999809265137,81            1.0299999713897705,82            1.0299999713897705,83            1.0499999523162842,84            1.0499999523162842,85            1.0499999523162842,86            1.0499999523162842,87            1.0499999523162842,88            1.0699999332427979,89            1.0999999046325684,90            1.1099998950958252,91            1.1599998474121094,92            1.1599998474121094,93            1.1699998378753662,94            1.2899998426437378,95            1.339999794960022,96            1.679999828338623,97            1.7899998426437378,98            1.8199998140335083,99            1.8499997854232788,100            1.8799997568130493,101            1.9099997282028198,102            1.9399996995925903,103            1.9899996519088745,104            2.0199997425079346,105            2.0199997425079346,106            2.0199997425079346,107            2.0199997425079346,108            2.0199997425079346,109            2.0199997425079346,110            2.0299997329711914,111            2.0299997329711914,112            2.0299997329711914,113            2.0299997329711914,114            2.0299997329711914,115            2.0299997329711914,116            2.0299997329711914,117            2.0299997329711914,118            2.0299997329711914,119            2.0799996852874756,120            2.0899996757507324,121            2.189999580383301,122            2.2199995517730713,123            2.5899994373321533,124            2.729999542236328,125            2.749999523162842,126            2.8399994373321533127        ],128        "type": "longrope"129    },130    "rope_theta": 10000.0,131    "sliding_window": 262144,132    "tie_word_embeddings": false,133    "torch_dtype": "bfloat16",134    "transformers_version": "4.43.3",135    "use_cache": true,136    "vocab_size": 32064137}